- Email: [email protected]
Flaps of the latissimus dorsi muscle in difficult wounds of the trunk and arm
Flaps of the Latissimus Dorsi Muscle in Difficult Wounds Of the Trunk and Arm
William D. Morain, MD, Hanover, New Hampshire
The latissimus dorsi muscle has the largest surface area of any extremity-related muscle unit of the human body. Thii fact and its dual blood supply give it special value as a tool in reconstructive surgery. It.3 location in the posterolateral thorax and the strategic position of its principal pivot axis in the axilla make it a uniquely versatile muscle for reconstruction of the abdomen, thorax, vertebral area, and arm [l--3]. The muscle takes an aponeurotic origin from the spinous processes of the lowermost seven thoracic vertebrae, the lumbar and sacral vertebrae, and the iliac crest [4]. The fibers insert as a broad, thick tendon in the intertubercular groove of the humerus. It acts as an adductor and medial rotator of the arm and draws the arm posteriorly as well. Its nerve supply is the thoracodorsal nerve from the posterior cord of the brachial plexus. The dominant vascular supply to the latissimus dorsi is the single thoracodorsal artery and vein which arise from the subscapular vessels and in turn the axillary vessels. Variations in this pedicle have been described by Bartlett et al [5]. Within the muscle itself, the thoracodorsal vessels bifurcate, each branch coursing more or less parallel to the muscle fibers and supplying the medial and lateral halves independently [S]. Musculocutaneous perforators penetrate the muscle to supply the adjacent overlying skin. The entire muscle with any portion of overlying skin may be rotated about the thoracodorsal vessel pivot point to reach anterior and posterior trunk and arm defects. A secondary blood supply exists as a series of large, paraspinous perforating vessels which arise parallel to and about 4 cm lateral to the vertebral spines [ 7,8]. These segmental
From the Section of Plastic and Reconstructive Surgery, Dartmouth-Hitchcock Medical Center, Hanover, New Hampshire. Requests for reprints should be addressed to William D. Morain. MD, Sectii of Plastic and Fteconstnxtiie Surgery, Dartmouth-Hi&cock Medical Center. Hanover. New Hamoshire 03755. Presented at the 63rd A&al Meeting of the New England Surgical Society, Bretton Woods, New Hampshire, October 15-17, 1982.
520
vessels course in a generally lateral direction and will carry sizable segmental portions of the latissimus muscle as a distally based flap along with its overlying skin if needed [9]. Case Reports Case 1. Reconstruction of scapular region: A 58 year old man with known metastatic adenocarcinoma of the prostate had been treated with 4,000 rads for a large plasmacytoid lymphoma of the subcutaneous tissues overlying the left scapula 3.5 years previously. A severe ulcerating dermatitis of the irradiated area resulted. A recurrence 6 months before admission had been treated with chemotherapy. On admission the patient showed widespread osteoradionecrosis of the left scapula. At surgery 80 percent of the scapula was resected, preserving only the lateral supraspinatus portion with the acromion, the glenoid cavity, and the coracoid process. The resulting 10 by 16 cm quadrilateral defect was reconstructed with a left latissimus dorsi myocutaneous flap with a tailored skin paddle and skin graft of the donor defect. Healing was uneventful and satisfactory shoulder function was preserved. Case 2. Reconstruction of sternocostal region: A 62 year old woman underwent a radical mastectomy of the left breast for an infiltrating ductal carcinoma. All axillary nodes were negative on microscopic examination. Four years later local recurrence developed in the left medial chest wall which was treated with 4,000 rads. Slow progression of the local recurrent disease with ulceration was noted over the next several years. At the age of 72 the patient presented with a 4 by 10 cm ulcerated mass in the left anteromedial chest wall which was adherent to the sternum and the fourth through seventh ribs, She underwent a full-thickness en bloc resection of the chest wall with removal of the anterior sections of four ribs and a portion of the sternum. Primary reconstruction was carried out with a left latissimus dorsi myocutaneous flap which healed completely. Bony stabilization was unnecessary. Satisfactory healing occurred with ample chest wall rigidity being provided by the dense latissimus muscle and fascia. Case 3. Reconstruction of anterior thoracoabdominal region: A 67 year old man had been treated 4 years previously for an adenocarcinoma of the transverse colon
The American Journal of Surgery
Managing Difficult Wounds of the Trunk and Arm
with abdominal wall penetration. Initial management consisted of administration of 4,500 rads followed by en bloc resection of the transverse and descending colon and a small full-thickness segment of the abdominal wall including the tumor and sinus tract. Primary closure of the abdominal wall was carried out. Two months before admission an incisional hernia in the wound area was explored, and unexpectedly, a large 15 by 20 by 8 cm recurrent tumor mass invading the adjacent abdominal wall and the lowest three ribs was found. Manual abdominal exploration was otherwise unremarkable. Marlexe mesh was used in the repair of the hernia. At the time of the patient’s admission a large, hard tumor mass on the anterior thoracoabdominal wall in the irradiated left upper quadrant was found. A thorough work-up for distant metastatic disease disclosed only an enlarged right adrenal gland on computerized tomographic (CT) scan. A right adrenalectomy was performed and disclosed an adenoma. Thereafter, a left anterior thoracoabdominal wall resection was carried out, removing as well a 10 cm segment of small intestine which was adherent to the Marlex mesh. Sections of three ribs and a portion of the diaphragm were removed in en bloc fashion with the specimen. Margins were free of tumor on microscopic examination. A flap of the adjacent rectus abdominus muscle was not possible because previous surgery had divided its superior axial blood supply. The 16 by 17 cm full-thickness defect was reconstructed with a left latissimus dorsi myocutaneous flap with a 22 by 16 cm skin paddle and a splitthickness skin graft to the donor site. Healing was uneventful and the patient was active with no evident tumor 1 year after the procedure. Case 4. Reconstruction of axilla and breast (Figure 1): A 40 year old woman had sustained flame burns to the right anterolateral chest wall at the age of 4 years. Her treatment had included pinch grafting and a prolonged period of secondary epithelization. She was left in adult life with a fixed adduction contracture of the right shoulder, hypoplasia, and distortion of the right breast. A wide area of stable, irregular scar was present over the upper anterolateral trunk. The right nipple remnant was at the level of the inframammary fold on the left side. Releases of both the axillary and breast contractures were carried out by curvilinear incisions in the anterior axillary fold and in the planned inframammary fold. A split latissimus myocutaneous flap was transferred with separate 6 by 12 cm crescentic skin paddles which were inserted into the respective defects (Figure 1). Placement of a breast implant was deferred because of residual scar tightness. Significant improvement in shoulder elevation and breast contour resulted. Case 5. Distally based flap in reconstruction of vertebral region: A 66 year old man presented with a recurrent intramedullary ependymoma of the thoracic spine. Two years previously he had undergone a laminectomy of the 8th through the 11th thoracic vertebrae with partial excision of the tumor, followed by the administration of 5,000 rads over 6 weeks. Spastic paraplegia developed but subsequently stabilized. With recurrence documented on admission by myelography at the Sth, 9th, and 10th thoracic vertebrae reexploration was carried out with resection of the tumor. A major dehiscence through all layers occurred 2 weeks after surgery and was associated with fever, cerebral spinal fluid
Volume 145, April 1983
Figure 1. The split right latissimus do& myocutaneous flap with a double skin paddle.
leak, and meningitis. An attempt at direct repair was unsuccessful. Three weeks later a second attempt at repair also failed and the cerebral spinal fluid leak continued. Seven weeks after resection a 9 by 19 cm distally based latissimus dorsi myocutaneous flap was transferred to the ischemic open wound in conjunction with the placement of a ventricular drain. Healing was uneventful and the cerebral spinal fluid leak was immediately sealed. Case 6. Reconstruction of massive vertebral defect: A 56 year old insulin-dependent diabetic man presented with a 15 year history of multiple recurrences of a large basal cell carcinoma of his mid-back. Two years previously he had undergone an incomplete resection at which time a tumor was found in the spinous processes of the 10th and 11th thoracic vertebrae and the 1st lumbar vertebra. T12 involvement of the 12th thoracic vertebra was radiologically suspicious as well. The patient had received 5,000 rads to the thoracolumbar spine at that time. A large ischemic Pseudomonas-laden ulcer resulted. Six months later he underwent debridement of the radiation-induced ulcer with the finding of widespread residual tumor. On this admission a CT scan showed widespread soft tissue tumor in the paraspinous region with a normal intraspinous configuration of epidural fat. En bloc resection of the tumor was carried out, removing a 12 by 21 cm skin segment, including all the regional paraspinous muscles in continuity with bilateral laminectomy of the eight vertebral segments from the seventh thoracic through the second lumbar vertebrae. A total right latissimus dorsi myocutaneous flap was transferred over the defect and secured over suction drains. A meshed skin graft covered the flap donor area. The flap did not adhere in the deeply grooved concavity of the laminectomy defect because of a pseudomonal infection. When major flap separation occurred 2 weeks later, salvage was attempted by securing the flap along the distal edge of the spinal structures. Loss of dural integrity ensued, followed by meningitis and a right
521
h&rain
Figure 2. Chest sinogram. Contrast material ‘she ws large dimensions of old empyema cav/ty.
parietal intracerebral hemorrhage. Intrathecal antibiosis was instituted. For salvage, an omental flap was elevated on the left gastroepiploic vessels and tunneled through the left flank. It was draped over the exposed spinal cord after debridement of the devascularized regions of dura. The latissimus flap was then sutured back over the omentum and the re-
Flgure 3. Method of cavity.
522
passage of muscle Into the empyema
mainder of the wound skin was grafted. The dural leak was sealed completely by this procedure, the omentum providing the mechanical flexibility to conform to the hollow of the defect. However, the patient never regained cortical function and died 2.5 months after the initial procedure. Postmortem examination showed a complete dural seal and no residual basal cell carcinoma. Case 7. Intrathoracic reconstruction (Figure 2): A 54 year old man presented with a large pleurocutaneous fistula of the left side. His history dated from over 2 years before when he had contracted pneumonia. A left empyema ensued and had been surgically drained posteriorly through the 11th rib by an Eloesser procedure. The large drainage fenestration remained open and the left lung remained partially collapsed. The patient suffered general malaise and anorexia and sustained a 20 pound weight loss over the 2 year period. An attempt at prosthetic coverage and chest tube suction was unsuccessful. Chest roentgenogram showed a large communicating cavity occupying about one third of the left hemithorax (Figure 2). At surgery, the surface of the entire chest cavity was freshened with a uterine curette inserted through the fenestration. The latissimus dorsi muscle was exposed in its entirety and divided at the sites of origin and insertion. It was passed into the cavity through an opening created by removal of segments of two ribs at the superior aspect of the defect, remaining attached only by the thoracodorsal vascular pedicle (Figure 3). The muscle flap was secured around the periphery of the defect and a separate local turnover flap of fascia was sutured over the outside of the fenestration. Four large suction drains were inserted to aid in lung expansion. The skin was closed directly. Primary healing occurred with roentgenographic confirmation of full pulmonary expansion. The patient regained weight and became fit for reemployment. Case 8. Reconstruction of the upper arm (Figure 4): A 31 year old right-handed man sustained a fall into a
The American Journal 01 Surgery
Managing Difficult Wounds of the Trunk and Arm
Figure 4. Case 6. Left, a right total latissimus dorsi myocutaneous flap being transferred to the upper arm. Right, the healed wound wffh restored radial nerve functions.
brush pile which caused an abrasion and small puncture wound of the right elbow. Within a few hours he was seen briefly in a local emergency room because of local swelling and pain. A roentgenogram of the arm was unremarkable, but his white blood cell count was 9,800 with 45 polymorphonuclear leukocytes, 46 band forms, and 1 metamyelocyte. He was discharged without specific treatment. Forty-eight hours after the injury he was admitted to Dartmouth-Hitchcock Medical Center because of dramatic swelling from the right wrist to the chest wall associated with anorexia and malaise. Vital signs included a temperature of 38.5’ C, blood pressure of lOO/SO mm Hg, and a pulse of I20 beats/min. The right arm was tensely swollen with numerous hemorrhagic bullae over the upper arm. A small puncture wound was observed over the olecranon. Radial nerve function was completely absent. Doppler examination was unremarkable, crepitus was absent, and no interstitial gas was seen on roentgenogram. The white leublood cell count was 2,400 with 3 polymorphonuclear kocytes, 55 band forms, and 29 metamyelocytes. Aspiration of the blister fluid and of the olecranon bursa disclosed cloudy red fluid which contained many gram-positive cocci on a smear. Culture of the fluid, as well as blood cultures, grew P-hemolytic streptococci of group A. Initial therapy included administration of intravenous
Volume 145, April 1983
penicillin 4,000,OOOunits every 4 hours and immediate arm fasciotomy with wide drainage of the olecranon bursa. After 24 hours the patient showed increasing toxicity with marked regional and systemic deterioration. A diagnosis of acute streptococcal gangrene was made. With the appearance of obvious myonecrosis of the triceps, a second operation was carried out with fasciotomy extended across the shoulder and down the involved right lateral chest and abdominal wall to the right knee. Most of the triceps was excised. Rapid pulmonary decompensation ensued with infiltrates on roentgenogram. Cardiac ventricular arrythmias, severe acidosis, and seizures followed. Continuous wound hemorrhage was noted associated with marked thrombocytopenia. Two days later another fluorescein-assisted radical debridement was carried out with removal of all of the triceps, most of the teres major and minor, and portions of the biceps, infraspinatus, deltoid, and latissimus dorsi muscles. Topical silver sulfadiazine dressings were instituted for control of the wound sepsis. Six days later final surface debridement was carried out, sacrificing 15 cm of necrotic radial nerve and much of the humeral periosteum and dorsal elbow capsule. A total right latissimus dorsi myocutaneous flap was transferred to resurface the posterior arm, and the remainder of the wound
523
Morain
was closed directly or with meshed split-thickness
skin grafts (Figure 4, left). Further systemic decompensation ensued, including renal failure, upper gastrointestinal hemorrhage, pulmonary failure, and acute cholecystitis. Therapy included hemodialysis, tracheostomy, and cholecystectomy. All wounds were finally closed and he was discharged from the hospital (Figure 4, right). During his hospitalization he had consumed 256 units of blood products, including 89 units of blood and 112 units of albumin. Subsequent tendon transfers to restore radial nerve function permitted rehabilitation of the dominant right hand.
Comments As with all contemporary reconstructive surgery, careful planning is essential and poor planning is the principal cause of failure. Knowing the axis, the arc, and the cutaneous limits of the latissimus dorsi myocutaneous flap, advance measurement of geometric skin needs and accurate preliminary marking will permit a precise functional and anatomic result. Certain technical points merit emphasis in the transfer of the latissimus dorsi unit. Since neither origin nor insertion plays any role in the blood supply of the flap, both should be divided in complex reconstructive situations. Not only are mobility and arc of rotation dramatically increased by this maneuver, but one avoids the powerful pull exerted by the innervated latissimus dorsi muscle with intact insertion which may promote separation from its site of surgical attachment. (This probably played a contributing role in the complication encountered in Case 6.) This technical feature is in contrast to the flap transfer of the corresponding pectoralis major muscle on the anterior trunk where the dominant vascular pedicle of the latter runs perpendicular to the muscle fibers rather than parallel to them [IO]. The special problem of the persistent pleurocutaneous fistula has been addressed in a variety of surgical and nonsurgical ways, most commonly by collapsing the chest wall against the lung. Since the lung retains a surprising potential for reexpansion, even after a prolonged period of partial collapse, a thin, well-vascularized muscle flap interposed between visceral and parietal pleura may be sufficient to effect adherence, as in Case 7. The inclusion of subcutaneous fat and dermis in a myocutaneous unit of the latissimus dorsi may therefore be superfluous in this instance. Furthermore, since a vascularized muscle surface is far more effective in controlling a bacterial load on a wound surface than is fat or dermis [ 111, pleural adherence is more likely when both pleural surfaces appose muscle. Passing the muscle through a narrow tunnel, as in Case 7, is also technically advantageous since it avoids the need for an open thoracotomy. The latissimus dorsi muscle has excellent twodimensional tensile strength. Thus, it usually will maintain sufficient rigidity in abdominal and moderate-size chest wall reconstructions so that allo-
524
plastic mesh or bone graft support is unnecessary. This is fortunate since the ulceration in many irradiated or malignant lesions (such as in Case 2) predisposes to a contaminated wound bed with a dramatically increased risk of infection in the presence of nonvascular materials. Summary The latissimus dorsi muscle is uniquely positioned for extraordinary utility as a reconstructive tool. Its versatility in transfer as a muscle or myocutaneous flap unit permits closure of a variety of difficult, complex wounds of the trunk and arm. Eight cases are presented which illustrate numerous variations of flap design for reconstructing large wounds of the scapular region, the sternocostal region, the anterior thoracoabdominal region, the axilla and breast, the vertebral and paravertebral regions, the intrathoracic cavities associated with pleurocutaneous fistulae, and the arm. The flap may be based on either the proximal or distal blood supply. It may be transferred in a single unit or split longitudinally. It may or may not carry overlying skin of any size or configuration. Transfer without overlying skin is advocated when the flap is placed intrathoracically for pleurocutaneous fistula closure. Acknowledgmenti I acknowledge the collaborative participation of Drs. Stephen Plume, Richard Saunders, Harry Goldsmith, and Arthur Naitove in the operative care of two of our patients (Cases 3 and 6). References 1. McGraw JB, Penix JO, Baker JW. Repair of major defects of the chest wall and spine with the latissimus dorsi myocutaneous flap. Plast Reconstr Surg 1978;62:197-206. 2. Bostwick J, Nahai F, Wallace JG, Vasconez LO. Sixty latissimus flaps. Plast Reconstr Surg 1979;63:31-41. 3. Shesol BF, Clarke JS. lntrathoracic application of the latissimus dorsi musculocutaneous flap. Plast Reconstr Surg 1980; 66:842-5. 4. Lockhart RD. Myology. In: Romanes 61, ed. Cunningham’s textbook of anatomy. 10th Ed. London: Oxford University Press, 1964:316. 5. Bartlett SP, May JW, Yaremchuk MJ. The latissimus dorsi muscle: a fresh cadaver study of the primary neurovascular pedicle. Plast Reconstr Surg 1981;67:631-6. 6. Tobin GR, Schusterman M, Peterson GH. Nichols G, Bland Kl. The intramuscular neurovascular anatomy of the latissimus dorsi muscle: the basis for splitting the flap. Plast Reconstr Surg 1981;67:637-41. 7. Manchot C. The cutaneous arteries of the human body. Translated by Ristic J, Morain W. New York: Springer-Verlag. 1983 (in press). 8. Bostwick J, Scheflan M, Nahai F, Jurkiewicz MJ. The “reverse” latissimus dorsi muscle and musculocutaneous flap: anatomical and clinical considerations. Plast Reconstr Surg 1980;65:395-9. 9. Nahai F, Scheflan M. Bostwick J. Posterior trunk reconstruction. Chapter 17. In: Mathes SJ, Nahai F. eds. Clinical applications for muscle and musculocutaneous flaps. St. Louis: C. V. Mosby, 1982:349-61. 10. Scheflan M, Bostwick J, Nahai F. Chest wall reconstruction-management of the difficult chest wound. Ann Plast
The
American Journal of Surgery
Managing Difficult Wounds of the Trunk and Arm
Surg 1982;8:122-31. 11. Mathes SJ, Alpert B.S. Chang N. Use of the muscle flap in chronic osteomyelitis: experimental and clinical correlation. Plast Reconstr Surg 1982;89:815-28.
Discussion J. Gordon Scannell (Boston, MA): I believe it is time to paraphrase Ambrose Pare and say, “We make the wound and the plastic surgeon closes it.” Some of our former members who dealt with difficult cancer problems with dreadful defects and excessive radiation therapy would have been delighted to hear this paper. The plastic surgeon is a sort of born-again anatomist and he must know all the fine print in Gray’s Anatomy. I have two questions. The latissimus dorsi, beyond being an adductor, inward rotator, and retractor of the humerus, is also the great muscle of coughing. When you cough, and if you watch yourself in the mirror sometime, you will see that a strap is set across your chest. Has there been any respiratory interference in your patients? Finally, the case where a pleural space was filled with a muscle flap did provide an ingenious way to avoid the mutilating effects of a thoracoplasty by obliterating dead space, but I would prefer decortication to bring that lung up directly when possible. The presence of a bronchopleural fistula would enormously complicate and probably preclude the flap procedure. John B. Constable (Boston, MA): I think that the world of the closing of defects has been transformed by the use of musculocutaneous or myodermal flaps, and plastic surgeons seek them everywhere. Although Dr. Morain left the skin on the latissimus dorsi in most of his patients, we have tended more toward the use of direct grafting of the muscle without skin. Particularly in the obese patient, the skin with its subcutaneous tissue may make it too bulky. I would appreciate his thought on whether this is not a valuable procedure. The principal or the most common use of this flap today is in breast reconstruction. I believe it provides the best way of replacing the missing breast tissue with no effect on function. Joseph Upton (Boston, MA): I would like to ask Dr. Morain whether he ever denervates the muscle itself. I know when we have used it as a pedicle flap, especially
Volume 145, April 1993
when we have utilized the stuff principle and maintained much of its length, we had patients who were very symptomatic. So, we now remove at least a 2 to 3 cm segment of nerve when the latissimus dorsi is to be used as a pedicle flap. William D. Morain (closing): With regard to functional deficit from the latissimus dorsi, this has really never been reported as a major problem, although it has always been pointed out that there might be difficulties in athletes, particularly tennis players and swimmers. There are at least six muscles that take over portions of the function of the latissimus dorsi when it is missing. The only comment that I have ever had from patients about function has to do with their climbing out of a chair with their hand on the back and pressing upward: they notice a little bit of deficit here, but it is surprising that you can take a muscle of this size and not have any greater deficit than this. I agree that this is not a procedure to be used in the presence of a bronchopleural fistula because of the bacteriologic consequences. Dr. Constable used the two terms musculocutaneous and myodermal rather than myocutaneous. He and I have already had a dispute in print on which term is correct. I have always said myocutaneous, and he feels that this is an entomologic bastardization of Greek and Latin, perhaps not realizing that so much of our language is already thus bastardized. We have tended to continue to use the overlying skin along with the latissimus dorsi muscle most of the time, the one exception being when it is transferred to the foot as a free flap, in which case it is much too bulky with theoverlying skin and should instead be skin grafted. We have continued keeping our flaps innervated. I think one of the possible problems that caused a major complication had to do with leaving the insertion intact in the patient with cancer of the mid-back (Case 6). I think it exerted too much pull. A possible problem is that a very strong innervated muscle may help the flap to separate. Denervation might correct that problem. Certainly, denervation should probably not be carried out when flaps are transferred to the head and neck region, because in this region it is very important that the suppleness of that muscle remains intact in order that contracture does not develop in the pharynx.
525
William D. Morain, MD, Hanover, New Hampshire
The latissimus dorsi muscle has the largest surface area of any extremity-related muscle unit of the human body. Thii fact and its dual blood supply give it special value as a tool in reconstructive surgery. It.3 location in the posterolateral thorax and the strategic position of its principal pivot axis in the axilla make it a uniquely versatile muscle for reconstruction of the abdomen, thorax, vertebral area, and arm [l--3]. The muscle takes an aponeurotic origin from the spinous processes of the lowermost seven thoracic vertebrae, the lumbar and sacral vertebrae, and the iliac crest [4]. The fibers insert as a broad, thick tendon in the intertubercular groove of the humerus. It acts as an adductor and medial rotator of the arm and draws the arm posteriorly as well. Its nerve supply is the thoracodorsal nerve from the posterior cord of the brachial plexus. The dominant vascular supply to the latissimus dorsi is the single thoracodorsal artery and vein which arise from the subscapular vessels and in turn the axillary vessels. Variations in this pedicle have been described by Bartlett et al [5]. Within the muscle itself, the thoracodorsal vessels bifurcate, each branch coursing more or less parallel to the muscle fibers and supplying the medial and lateral halves independently [S]. Musculocutaneous perforators penetrate the muscle to supply the adjacent overlying skin. The entire muscle with any portion of overlying skin may be rotated about the thoracodorsal vessel pivot point to reach anterior and posterior trunk and arm defects. A secondary blood supply exists as a series of large, paraspinous perforating vessels which arise parallel to and about 4 cm lateral to the vertebral spines [ 7,8]. These segmental
From the Section of Plastic and Reconstructive Surgery, Dartmouth-Hitchcock Medical Center, Hanover, New Hampshire. Requests for reprints should be addressed to William D. Morain. MD, Sectii of Plastic and Fteconstnxtiie Surgery, Dartmouth-Hi&cock Medical Center. Hanover. New Hamoshire 03755. Presented at the 63rd A&al Meeting of the New England Surgical Society, Bretton Woods, New Hampshire, October 15-17, 1982.
520
vessels course in a generally lateral direction and will carry sizable segmental portions of the latissimus muscle as a distally based flap along with its overlying skin if needed [9]. Case Reports Case 1. Reconstruction of scapular region: A 58 year old man with known metastatic adenocarcinoma of the prostate had been treated with 4,000 rads for a large plasmacytoid lymphoma of the subcutaneous tissues overlying the left scapula 3.5 years previously. A severe ulcerating dermatitis of the irradiated area resulted. A recurrence 6 months before admission had been treated with chemotherapy. On admission the patient showed widespread osteoradionecrosis of the left scapula. At surgery 80 percent of the scapula was resected, preserving only the lateral supraspinatus portion with the acromion, the glenoid cavity, and the coracoid process. The resulting 10 by 16 cm quadrilateral defect was reconstructed with a left latissimus dorsi myocutaneous flap with a tailored skin paddle and skin graft of the donor defect. Healing was uneventful and satisfactory shoulder function was preserved. Case 2. Reconstruction of sternocostal region: A 62 year old woman underwent a radical mastectomy of the left breast for an infiltrating ductal carcinoma. All axillary nodes were negative on microscopic examination. Four years later local recurrence developed in the left medial chest wall which was treated with 4,000 rads. Slow progression of the local recurrent disease with ulceration was noted over the next several years. At the age of 72 the patient presented with a 4 by 10 cm ulcerated mass in the left anteromedial chest wall which was adherent to the sternum and the fourth through seventh ribs, She underwent a full-thickness en bloc resection of the chest wall with removal of the anterior sections of four ribs and a portion of the sternum. Primary reconstruction was carried out with a left latissimus dorsi myocutaneous flap which healed completely. Bony stabilization was unnecessary. Satisfactory healing occurred with ample chest wall rigidity being provided by the dense latissimus muscle and fascia. Case 3. Reconstruction of anterior thoracoabdominal region: A 67 year old man had been treated 4 years previously for an adenocarcinoma of the transverse colon
The American Journal of Surgery
Managing Difficult Wounds of the Trunk and Arm
with abdominal wall penetration. Initial management consisted of administration of 4,500 rads followed by en bloc resection of the transverse and descending colon and a small full-thickness segment of the abdominal wall including the tumor and sinus tract. Primary closure of the abdominal wall was carried out. Two months before admission an incisional hernia in the wound area was explored, and unexpectedly, a large 15 by 20 by 8 cm recurrent tumor mass invading the adjacent abdominal wall and the lowest three ribs was found. Manual abdominal exploration was otherwise unremarkable. Marlexe mesh was used in the repair of the hernia. At the time of the patient’s admission a large, hard tumor mass on the anterior thoracoabdominal wall in the irradiated left upper quadrant was found. A thorough work-up for distant metastatic disease disclosed only an enlarged right adrenal gland on computerized tomographic (CT) scan. A right adrenalectomy was performed and disclosed an adenoma. Thereafter, a left anterior thoracoabdominal wall resection was carried out, removing as well a 10 cm segment of small intestine which was adherent to the Marlex mesh. Sections of three ribs and a portion of the diaphragm were removed in en bloc fashion with the specimen. Margins were free of tumor on microscopic examination. A flap of the adjacent rectus abdominus muscle was not possible because previous surgery had divided its superior axial blood supply. The 16 by 17 cm full-thickness defect was reconstructed with a left latissimus dorsi myocutaneous flap with a 22 by 16 cm skin paddle and a splitthickness skin graft to the donor site. Healing was uneventful and the patient was active with no evident tumor 1 year after the procedure. Case 4. Reconstruction of axilla and breast (Figure 1): A 40 year old woman had sustained flame burns to the right anterolateral chest wall at the age of 4 years. Her treatment had included pinch grafting and a prolonged period of secondary epithelization. She was left in adult life with a fixed adduction contracture of the right shoulder, hypoplasia, and distortion of the right breast. A wide area of stable, irregular scar was present over the upper anterolateral trunk. The right nipple remnant was at the level of the inframammary fold on the left side. Releases of both the axillary and breast contractures were carried out by curvilinear incisions in the anterior axillary fold and in the planned inframammary fold. A split latissimus myocutaneous flap was transferred with separate 6 by 12 cm crescentic skin paddles which were inserted into the respective defects (Figure 1). Placement of a breast implant was deferred because of residual scar tightness. Significant improvement in shoulder elevation and breast contour resulted. Case 5. Distally based flap in reconstruction of vertebral region: A 66 year old man presented with a recurrent intramedullary ependymoma of the thoracic spine. Two years previously he had undergone a laminectomy of the 8th through the 11th thoracic vertebrae with partial excision of the tumor, followed by the administration of 5,000 rads over 6 weeks. Spastic paraplegia developed but subsequently stabilized. With recurrence documented on admission by myelography at the Sth, 9th, and 10th thoracic vertebrae reexploration was carried out with resection of the tumor. A major dehiscence through all layers occurred 2 weeks after surgery and was associated with fever, cerebral spinal fluid
Volume 145, April 1983
Figure 1. The split right latissimus do& myocutaneous flap with a double skin paddle.
leak, and meningitis. An attempt at direct repair was unsuccessful. Three weeks later a second attempt at repair also failed and the cerebral spinal fluid leak continued. Seven weeks after resection a 9 by 19 cm distally based latissimus dorsi myocutaneous flap was transferred to the ischemic open wound in conjunction with the placement of a ventricular drain. Healing was uneventful and the cerebral spinal fluid leak was immediately sealed. Case 6. Reconstruction of massive vertebral defect: A 56 year old insulin-dependent diabetic man presented with a 15 year history of multiple recurrences of a large basal cell carcinoma of his mid-back. Two years previously he had undergone an incomplete resection at which time a tumor was found in the spinous processes of the 10th and 11th thoracic vertebrae and the 1st lumbar vertebra. T12 involvement of the 12th thoracic vertebra was radiologically suspicious as well. The patient had received 5,000 rads to the thoracolumbar spine at that time. A large ischemic Pseudomonas-laden ulcer resulted. Six months later he underwent debridement of the radiation-induced ulcer with the finding of widespread residual tumor. On this admission a CT scan showed widespread soft tissue tumor in the paraspinous region with a normal intraspinous configuration of epidural fat. En bloc resection of the tumor was carried out, removing a 12 by 21 cm skin segment, including all the regional paraspinous muscles in continuity with bilateral laminectomy of the eight vertebral segments from the seventh thoracic through the second lumbar vertebrae. A total right latissimus dorsi myocutaneous flap was transferred over the defect and secured over suction drains. A meshed skin graft covered the flap donor area. The flap did not adhere in the deeply grooved concavity of the laminectomy defect because of a pseudomonal infection. When major flap separation occurred 2 weeks later, salvage was attempted by securing the flap along the distal edge of the spinal structures. Loss of dural integrity ensued, followed by meningitis and a right
521
h&rain
Figure 2. Chest sinogram. Contrast material ‘she ws large dimensions of old empyema cav/ty.
parietal intracerebral hemorrhage. Intrathecal antibiosis was instituted. For salvage, an omental flap was elevated on the left gastroepiploic vessels and tunneled through the left flank. It was draped over the exposed spinal cord after debridement of the devascularized regions of dura. The latissimus flap was then sutured back over the omentum and the re-
Flgure 3. Method of cavity.
522
passage of muscle Into the empyema
mainder of the wound skin was grafted. The dural leak was sealed completely by this procedure, the omentum providing the mechanical flexibility to conform to the hollow of the defect. However, the patient never regained cortical function and died 2.5 months after the initial procedure. Postmortem examination showed a complete dural seal and no residual basal cell carcinoma. Case 7. Intrathoracic reconstruction (Figure 2): A 54 year old man presented with a large pleurocutaneous fistula of the left side. His history dated from over 2 years before when he had contracted pneumonia. A left empyema ensued and had been surgically drained posteriorly through the 11th rib by an Eloesser procedure. The large drainage fenestration remained open and the left lung remained partially collapsed. The patient suffered general malaise and anorexia and sustained a 20 pound weight loss over the 2 year period. An attempt at prosthetic coverage and chest tube suction was unsuccessful. Chest roentgenogram showed a large communicating cavity occupying about one third of the left hemithorax (Figure 2). At surgery, the surface of the entire chest cavity was freshened with a uterine curette inserted through the fenestration. The latissimus dorsi muscle was exposed in its entirety and divided at the sites of origin and insertion. It was passed into the cavity through an opening created by removal of segments of two ribs at the superior aspect of the defect, remaining attached only by the thoracodorsal vascular pedicle (Figure 3). The muscle flap was secured around the periphery of the defect and a separate local turnover flap of fascia was sutured over the outside of the fenestration. Four large suction drains were inserted to aid in lung expansion. The skin was closed directly. Primary healing occurred with roentgenographic confirmation of full pulmonary expansion. The patient regained weight and became fit for reemployment. Case 8. Reconstruction of the upper arm (Figure 4): A 31 year old right-handed man sustained a fall into a
The American Journal 01 Surgery
Managing Difficult Wounds of the Trunk and Arm
Figure 4. Case 6. Left, a right total latissimus dorsi myocutaneous flap being transferred to the upper arm. Right, the healed wound wffh restored radial nerve functions.
brush pile which caused an abrasion and small puncture wound of the right elbow. Within a few hours he was seen briefly in a local emergency room because of local swelling and pain. A roentgenogram of the arm was unremarkable, but his white blood cell count was 9,800 with 45 polymorphonuclear leukocytes, 46 band forms, and 1 metamyelocyte. He was discharged without specific treatment. Forty-eight hours after the injury he was admitted to Dartmouth-Hitchcock Medical Center because of dramatic swelling from the right wrist to the chest wall associated with anorexia and malaise. Vital signs included a temperature of 38.5’ C, blood pressure of lOO/SO mm Hg, and a pulse of I20 beats/min. The right arm was tensely swollen with numerous hemorrhagic bullae over the upper arm. A small puncture wound was observed over the olecranon. Radial nerve function was completely absent. Doppler examination was unremarkable, crepitus was absent, and no interstitial gas was seen on roentgenogram. The white leublood cell count was 2,400 with 3 polymorphonuclear kocytes, 55 band forms, and 29 metamyelocytes. Aspiration of the blister fluid and of the olecranon bursa disclosed cloudy red fluid which contained many gram-positive cocci on a smear. Culture of the fluid, as well as blood cultures, grew P-hemolytic streptococci of group A. Initial therapy included administration of intravenous
Volume 145, April 1983
penicillin 4,000,OOOunits every 4 hours and immediate arm fasciotomy with wide drainage of the olecranon bursa. After 24 hours the patient showed increasing toxicity with marked regional and systemic deterioration. A diagnosis of acute streptococcal gangrene was made. With the appearance of obvious myonecrosis of the triceps, a second operation was carried out with fasciotomy extended across the shoulder and down the involved right lateral chest and abdominal wall to the right knee. Most of the triceps was excised. Rapid pulmonary decompensation ensued with infiltrates on roentgenogram. Cardiac ventricular arrythmias, severe acidosis, and seizures followed. Continuous wound hemorrhage was noted associated with marked thrombocytopenia. Two days later another fluorescein-assisted radical debridement was carried out with removal of all of the triceps, most of the teres major and minor, and portions of the biceps, infraspinatus, deltoid, and latissimus dorsi muscles. Topical silver sulfadiazine dressings were instituted for control of the wound sepsis. Six days later final surface debridement was carried out, sacrificing 15 cm of necrotic radial nerve and much of the humeral periosteum and dorsal elbow capsule. A total right latissimus dorsi myocutaneous flap was transferred to resurface the posterior arm, and the remainder of the wound
523
Morain
was closed directly or with meshed split-thickness
skin grafts (Figure 4, left). Further systemic decompensation ensued, including renal failure, upper gastrointestinal hemorrhage, pulmonary failure, and acute cholecystitis. Therapy included hemodialysis, tracheostomy, and cholecystectomy. All wounds were finally closed and he was discharged from the hospital (Figure 4, right). During his hospitalization he had consumed 256 units of blood products, including 89 units of blood and 112 units of albumin. Subsequent tendon transfers to restore radial nerve function permitted rehabilitation of the dominant right hand.
Comments As with all contemporary reconstructive surgery, careful planning is essential and poor planning is the principal cause of failure. Knowing the axis, the arc, and the cutaneous limits of the latissimus dorsi myocutaneous flap, advance measurement of geometric skin needs and accurate preliminary marking will permit a precise functional and anatomic result. Certain technical points merit emphasis in the transfer of the latissimus dorsi unit. Since neither origin nor insertion plays any role in the blood supply of the flap, both should be divided in complex reconstructive situations. Not only are mobility and arc of rotation dramatically increased by this maneuver, but one avoids the powerful pull exerted by the innervated latissimus dorsi muscle with intact insertion which may promote separation from its site of surgical attachment. (This probably played a contributing role in the complication encountered in Case 6.) This technical feature is in contrast to the flap transfer of the corresponding pectoralis major muscle on the anterior trunk where the dominant vascular pedicle of the latter runs perpendicular to the muscle fibers rather than parallel to them [IO]. The special problem of the persistent pleurocutaneous fistula has been addressed in a variety of surgical and nonsurgical ways, most commonly by collapsing the chest wall against the lung. Since the lung retains a surprising potential for reexpansion, even after a prolonged period of partial collapse, a thin, well-vascularized muscle flap interposed between visceral and parietal pleura may be sufficient to effect adherence, as in Case 7. The inclusion of subcutaneous fat and dermis in a myocutaneous unit of the latissimus dorsi may therefore be superfluous in this instance. Furthermore, since a vascularized muscle surface is far more effective in controlling a bacterial load on a wound surface than is fat or dermis [ 111, pleural adherence is more likely when both pleural surfaces appose muscle. Passing the muscle through a narrow tunnel, as in Case 7, is also technically advantageous since it avoids the need for an open thoracotomy. The latissimus dorsi muscle has excellent twodimensional tensile strength. Thus, it usually will maintain sufficient rigidity in abdominal and moderate-size chest wall reconstructions so that allo-
524
plastic mesh or bone graft support is unnecessary. This is fortunate since the ulceration in many irradiated or malignant lesions (such as in Case 2) predisposes to a contaminated wound bed with a dramatically increased risk of infection in the presence of nonvascular materials. Summary The latissimus dorsi muscle is uniquely positioned for extraordinary utility as a reconstructive tool. Its versatility in transfer as a muscle or myocutaneous flap unit permits closure of a variety of difficult, complex wounds of the trunk and arm. Eight cases are presented which illustrate numerous variations of flap design for reconstructing large wounds of the scapular region, the sternocostal region, the anterior thoracoabdominal region, the axilla and breast, the vertebral and paravertebral regions, the intrathoracic cavities associated with pleurocutaneous fistulae, and the arm. The flap may be based on either the proximal or distal blood supply. It may be transferred in a single unit or split longitudinally. It may or may not carry overlying skin of any size or configuration. Transfer without overlying skin is advocated when the flap is placed intrathoracically for pleurocutaneous fistula closure. Acknowledgmenti I acknowledge the collaborative participation of Drs. Stephen Plume, Richard Saunders, Harry Goldsmith, and Arthur Naitove in the operative care of two of our patients (Cases 3 and 6). References 1. McGraw JB, Penix JO, Baker JW. Repair of major defects of the chest wall and spine with the latissimus dorsi myocutaneous flap. Plast Reconstr Surg 1978;62:197-206. 2. Bostwick J, Nahai F, Wallace JG, Vasconez LO. Sixty latissimus flaps. Plast Reconstr Surg 1979;63:31-41. 3. Shesol BF, Clarke JS. lntrathoracic application of the latissimus dorsi musculocutaneous flap. Plast Reconstr Surg 1980; 66:842-5. 4. Lockhart RD. Myology. In: Romanes 61, ed. Cunningham’s textbook of anatomy. 10th Ed. London: Oxford University Press, 1964:316. 5. Bartlett SP, May JW, Yaremchuk MJ. The latissimus dorsi muscle: a fresh cadaver study of the primary neurovascular pedicle. Plast Reconstr Surg 1981;67:631-6. 6. Tobin GR, Schusterman M, Peterson GH. Nichols G, Bland Kl. The intramuscular neurovascular anatomy of the latissimus dorsi muscle: the basis for splitting the flap. Plast Reconstr Surg 1981;67:637-41. 7. Manchot C. The cutaneous arteries of the human body. Translated by Ristic J, Morain W. New York: Springer-Verlag. 1983 (in press). 8. Bostwick J, Scheflan M, Nahai F, Jurkiewicz MJ. The “reverse” latissimus dorsi muscle and musculocutaneous flap: anatomical and clinical considerations. Plast Reconstr Surg 1980;65:395-9. 9. Nahai F, Scheflan M. Bostwick J. Posterior trunk reconstruction. Chapter 17. In: Mathes SJ, Nahai F. eds. Clinical applications for muscle and musculocutaneous flaps. St. Louis: C. V. Mosby, 1982:349-61. 10. Scheflan M, Bostwick J, Nahai F. Chest wall reconstruction-management of the difficult chest wound. Ann Plast
The
American Journal of Surgery
Managing Difficult Wounds of the Trunk and Arm
Surg 1982;8:122-31. 11. Mathes SJ, Alpert B.S. Chang N. Use of the muscle flap in chronic osteomyelitis: experimental and clinical correlation. Plast Reconstr Surg 1982;89:815-28.
Discussion J. Gordon Scannell (Boston, MA): I believe it is time to paraphrase Ambrose Pare and say, “We make the wound and the plastic surgeon closes it.” Some of our former members who dealt with difficult cancer problems with dreadful defects and excessive radiation therapy would have been delighted to hear this paper. The plastic surgeon is a sort of born-again anatomist and he must know all the fine print in Gray’s Anatomy. I have two questions. The latissimus dorsi, beyond being an adductor, inward rotator, and retractor of the humerus, is also the great muscle of coughing. When you cough, and if you watch yourself in the mirror sometime, you will see that a strap is set across your chest. Has there been any respiratory interference in your patients? Finally, the case where a pleural space was filled with a muscle flap did provide an ingenious way to avoid the mutilating effects of a thoracoplasty by obliterating dead space, but I would prefer decortication to bring that lung up directly when possible. The presence of a bronchopleural fistula would enormously complicate and probably preclude the flap procedure. John B. Constable (Boston, MA): I think that the world of the closing of defects has been transformed by the use of musculocutaneous or myodermal flaps, and plastic surgeons seek them everywhere. Although Dr. Morain left the skin on the latissimus dorsi in most of his patients, we have tended more toward the use of direct grafting of the muscle without skin. Particularly in the obese patient, the skin with its subcutaneous tissue may make it too bulky. I would appreciate his thought on whether this is not a valuable procedure. The principal or the most common use of this flap today is in breast reconstruction. I believe it provides the best way of replacing the missing breast tissue with no effect on function. Joseph Upton (Boston, MA): I would like to ask Dr. Morain whether he ever denervates the muscle itself. I know when we have used it as a pedicle flap, especially
Volume 145, April 1993
when we have utilized the stuff principle and maintained much of its length, we had patients who were very symptomatic. So, we now remove at least a 2 to 3 cm segment of nerve when the latissimus dorsi is to be used as a pedicle flap. William D. Morain (closing): With regard to functional deficit from the latissimus dorsi, this has really never been reported as a major problem, although it has always been pointed out that there might be difficulties in athletes, particularly tennis players and swimmers. There are at least six muscles that take over portions of the function of the latissimus dorsi when it is missing. The only comment that I have ever had from patients about function has to do with their climbing out of a chair with their hand on the back and pressing upward: they notice a little bit of deficit here, but it is surprising that you can take a muscle of this size and not have any greater deficit than this. I agree that this is not a procedure to be used in the presence of a bronchopleural fistula because of the bacteriologic consequences. Dr. Constable used the two terms musculocutaneous and myodermal rather than myocutaneous. He and I have already had a dispute in print on which term is correct. I have always said myocutaneous, and he feels that this is an entomologic bastardization of Greek and Latin, perhaps not realizing that so much of our language is already thus bastardized. We have tended to continue to use the overlying skin along with the latissimus dorsi muscle most of the time, the one exception being when it is transferred to the foot as a free flap, in which case it is much too bulky with theoverlying skin and should instead be skin grafted. We have continued keeping our flaps innervated. I think one of the possible problems that caused a major complication had to do with leaving the insertion intact in the patient with cancer of the mid-back (Case 6). I think it exerted too much pull. A possible problem is that a very strong innervated muscle may help the flap to separate. Denervation might correct that problem. Certainly, denervation should probably not be carried out when flaps are transferred to the head and neck region, because in this region it is very important that the suppleness of that muscle remains intact in order that contracture does not develop in the pharynx.
525