- Email: [email protected]
Groin lymphatic complications after arterial reconstruction
Groin lymphatic complications after arterial reconstruction Steve H. Tyndall, MD, Alexander D. Shepard, MD, Judith M. Wtlczewski, RN, BSN, Daniel J. Reddy, MD, Joseph P. Elliott, Jr., MD, and Calvin B. Ernst, MD, Detroit, Mich. Purpose: The purpose of this study was to better define the associated risks and optimal management of groin lymphatic complications (GLC) after femoral artery reconstructive operations. Methods: Retrospective review of a vascular surgery registry for the last 15 years identified 2679 arterial operations requiring a groin incision. Forty-one GLC were recognized, 28 lymphocutaneous fistulas (LF) and 13 lymphoceles. Results: The incidence of GLC was 1.5% per patient or 1.2% per incision. The highest incidence of GLC was in patients having an aortobifemoral bypass for aneurysmal disease in a previously operated groin (8.1% per patient) and in those undergoing an isolated femoral procedure in a previously operated groin (5.3%). The lowest frequency of GLC was after femoropopliteal/tibial bypasses (0.5%). Twenty-nine patients (71%) were treated without operation with bedrest, intravenous antibiotics, and aggressive local wound care. Operative therapy with wound reexploration attempted identification and control of the leak site, and meticulous wound closure was used in 12 patients (29%). Lymph fistulas in patients undergoing reoperation (10/28) resolved sooner than in patients treated without operation (18/28) (9 ± 3 days vs 24 ± 3 days). Infectious wound complications with one resultant graft infection developed in five of 18 patients with LF who did not undergo reoperation. There were no wound or graft infections in the patients in the LF group treated with operation. Operative exploration of lymphoceles did not reduce hospital stay or infectious wound complications. Repetitive lymphocele aspiration did not affect rapidity of resolution or increase the infectious complications. Conclusion: GLC remain a troublesome complication of femoral arterial reconstruction. Early reoperation should be performed once a LF is diagnosed. Treatment for lymphoceles should be individualized, with neither operative nor nonoperative management showing clear superiority. (J VASe SURG 1994;19:858-64.)
Groin lymphatic complications (GLC) are a rare but potentially serious problem after femoral arterial reconstruction. The risk of wound infection and subsequent graft infection, particularly of a synthetic vascular prosthesis, is a major concern. The cause of a lymphocutaneous fistula or lymphocele is presumed to be lymphatic disruption during groin dissection, but to date scant data are available to define risk factors. In addition controversy continues regarding optimal treatment, with some authorities favoring From the Department of Surgery, Division of Vascular Surgery, Henry Ford Hospital, Detroit. Presented at the Seventeenth Annual Meeting of the Midwestern Vascular Surgical Society, Chicago, Ill., Sept. 10-11, 1993. Reprint requests: Alexander D. Shepard, MD, Henry Ford Hospital, 2799 W. Grand Blvd., Detroit, MI 48202. Copyright © 1994 by The Society for Vascular Surgery and International Sociery for Cardiovascular Surgery, North American Chapter. 0741-5214/94/$3.00 + 0 24/6/53674
858
nonoperative therapyl,2 and others advocating early reoperation and direct control of the leak. 3,4 This review was undertaken in an effort to clarify these questions. MATERIAL AND METHODS The database of the Vascular Surgery Registry of the Henry Ford Hospital from 1977 through 1992 was retrospectively reviewed, and 41 patients with groin lymphatic complications after femoral arterial operations were identified. Data collected included patient demographics, operative indications, operation performed, presence of previous groin procedures, groin lymphatic complication treatment method, postoperative wound complications, and length of hospitalization and treatment. After arterial reconstruction all groin wounds were closed in three layers with interrupted absorbable sutures by either a senior resident or a staff
JOURNAL OF VASCULAR SURGERY Volume 19, Number 5
surgeon. Skin closure was based on surgeon preference and included staples, interrupted nylon stitches, or a running subcuticular absorbable suture. Lymphocutaneous fistulas were diagnosed by the appearance of persistent and near continuous clear drainage from a localized area of the groin wound most frequently near the distal end. Conservative treatment consisted of strict bed rest, limb elevation, prophylactic intravenous antibiotics, and aggressive local wound care, which included attempts at skin sterilization with povidone-iodine (Betadine) solution and pressure dressings. Operative therapy involved limited groin wound exploration. After removal of skin sutures and the first layer of subcutaneous stitches at the fistula site, draining fluid was tracked into the subcutaneous tissues, and an attempt was made to identify a leak site. Various vital dyes were injected into the first web space of the foot with subsequent calf/thigh compression in an attempt to promote operative identification. Presumed leak sites were suture ligated with absorbable suture. Exploration of the entire groin wound and removal of deep sutures closing tissue over the femoral artery or graft was avoided. After control of the presumed leaking lymphatic, the wound was meticulously closed. The choice of early or late wound exploration versus conservative management of the groin lymphatic complication was determined by individual staff surgeons. Lymphoceles were diagnosed by the presence of a persistent subcutaneous fluid collection in the absence of hematoma or wound abscess. These were either treated conservatively (bed rest, limited ambulation, or needle aspiration and pressure dressings) or by operative exploration, with excision of the lymphocele wall, attempted identification and control of the leak site, and meticulous wound closure. Data were statistically analyzed with the Student t test and Fisher's exact test. Data are expressed as the mean ± standard error. RESULTS Over the last 15 years a total of 2679 arterial reconstructive procedures requiring groin incisions were performed on the Vascular Surgery Service at the Henry Ford Hospital. Forty-one groin lymphatic complications consisting of 28 lymphocutaneous fistulas (68%) and 13 lymphoceles (32%) were recognized for an incidence of 1.5% per patient or 1.2% per groin incision. Of the 41 patients, 24 were men and 17 were women with an average age of 66 years (range 48 to 85 years). These sex and age distributions were not significantly different from
Tyndall et al.
859
those of patients undergoing femoral artery procedures in whom a GLC did not develop (Table I). Lymph complications were equally divided between the two groins, 21 on the right and 20 on the left. Mean follow-up after resolution of the groin lymphatic complication was 41 months (range 2 to 150 months). The percentage of redo groin operations, emergency procedures, and operations for tissue loss was no different for patients with and patients without lymphatic complications (Table I). Groin lymphatic problems occurred more frequently after aortobifemoral bypass (ABF) and isolated groin procedures (i.e., femoral anastomotic pseudoaneurysm repair, femoral endarterectomy, profundaplasty, graft thrombectomy, or embolectomy) than after infrainguinal bypasses (Table I). Although there was no overall difference in the incidence of groin lymphatic complications between patients with and without previous groin procedures, redo groin dissection did become a significant risk factor when analyzing specific operations. Patients undergoing an isolated femoral operation as a first groin procedure had a groin lymphatic complication incidence of 1.1 %, whereas those having an isolated femoral artery procedure after a previous groin operation had a 5.3% incidence (Table II). Similarly patients undergoing ABF grafting for abdominal aortic aneurysm repair had a groin lymphatic complication incidence of 0.7% in a nonoperated groin and 4.2% in a previously dissected groin (Table II). The lowest rate of groin lymphatic complications occurred in patients having femoropopliteal-tibial bypass grafting (0.5%). There was no difference in this group between patients with a previous femoral operation (0.3%) and those without one (0.7%) (Table II). When comparing patients with lymphocutaneous fistulas and lymphoceles, there was a significant difference in the number of days it took to diagnose the problem, 10 days for a lymphocutaneous fistula versus 18 days for a lymphocele (p = 0.02) (Table III). Most lymphocutaneous fistulas were diagnosed while patients were still hospitalized recovering from their reconstructive procedures (71 % ), whereas many lymphoceles were not diagnosed until the first outpatient visit (54%). Once a groin lymphatic complication was diagnosed, it took significantly longer for a lymphocele to resolve (37 ± 6 days) than it did for a lymphocutaneous fistula (20 ± 3 days) (p = 0.02). Five of eight patients diagnosed with a lymphocutaneous fistula after discharge from the hospital were readmitted for treatment, but only one
860
JOURNAL OF VASCULAR SURGERY May 1994
Tyndall et ai.
Table I. GLC: Patient demographics GLC (n = 41) Average age (yrs) Female Previous groin incision Tissue loss Emergency operation Operation performed Aortobifemoral graft Isolated femoral operation Femorodistal bypass
No GLC (n
= 2638)
P Value
66 (29%) (44%) (15%) (7%)
64 905 (34%) 1020 (39%) 524 (20%) 232 (9%)
NS NS NS NS NS
19 (46%) 15 (37%) 7 (17%)
665 (25%) 537 (20%) 1436 (54%)
<0.01 0.02 <0.01
12 18 6 3
Table II. Procedure-related incidence of GLC: Initial groin operations versus redo groin operations Initial groin operation
Redo groin operation
Per patient
Per incision
Per patient
Per incision
Operation (n)
(%)
(%)
(%)
(%)
P Value
All Operations (2679) APB (684) Isolated femoral (552) Femorodistal bypass (1443) APB for AAA (245) APB for occlusive disease (439)
1.4 2.7 1.1 0.7 1.4 3.5
1.0 1.4 1.1 0.7 0.7 1.8
1.7 3.0 5.3 0.3 8.1 1.5
1.5 1.5 5.3 0.3 4.2 0.8
NS NS <0.01
NS <0.05
NS
AAA, Abdominal aortic aneurysm.
Table III. Lymphocutaneous fistulas versus lymphoceles Interpal (mean days)
LF (n = 28)
Lymphocele (n = 13)
P Value
Initial operation to diagnosis Diagnosis to resolution Initial hospital stay Total hospital stay Infectious complications
10 ± 1 20 ± 3 16 ± 3 22 ± 2 18%
18 ± 3 37 ± 6 20 ± 2 18 ± 3 23%
0.02 0.02
of seven patients with a lymphocele discovered after discharge was rehospitalized. The interval from the original operation to resolution was significantly longer for a lymphocele (55 ± 9 days) than a lymphocutaneous fistula (30 ± 3 days). Skin closures at the time of the initial procedures in the 28 patients in whom a lymphocutaneous fistula developed included stapling in 13, interrupted nylon suturing in 13, and subcuticular absorbable suturing in two. Of the 13 patients in whom a lymphocele developed, five wounds were closed with staples, and eight were closed with interrupted nylon. Data on the method of groin wound closure was not available for patients without lymphatic complications. In general, however, staples and nylon sutures are used more frequently for groin wounds on our service than subcuticular closure. Eighteen patients with a lymphocutaneous fistula were treated without operation, and 10 underwent
NS NS NS
operative repair. There was no difference in the interval between the initial operation and the recognition of the lymph leak in these two groups (Table IV). Operative ligation of the leak site was successful in nine of 10 patients. At the time of reoperation, seven of these 10 patients had a leak site identified and ligated. In the three patients in whom no definite leak site was identified, a transected lymph node was found, though in no situation was it believed to be the definite source of the lymph leak. In two of these three patients the lymph drainage ceased within 24 hours of operation. In the other patient, lymph . drainage recurred rapidly after operation and required 24 days of conservative treatment before resolution, the average time it took for resolution in the nonoperative group. Injection of vital dyes did not help in the operative identification of leak sites. All identified lymph leaks were in the subcutaneous tissue. No graft or femoral artery was exposed in any
JOURNAL OF VASCULAR SURGERY
Tyndall et at.
Volume 19, Number 5
861
Table IV. Lymphocutaneous fistulas: Operative versus nonoperative therapy Interval (days) Initial operation to diagnosis Diagnosis to resolution Initial hospital stay Total hospital stay Infectious complications Graft infections
Operati-Pe (n = 10) 10 9 15 17
± ± ± ± 0 0
1 3 5 2
Nonopemti-Pe (n = 18)
P Value
10 ± 2 24 ± 3 23 ± 3 25 ± 3 5 (18%) 1 (6%)
0.84 <0.01 0.02 0.04 0.13 0.99
Nonaspirated (n = 6)
P Value
Table V. Lymphocele: Aspiration versus nonaspiration therapy Interval (mean days) Initial operation to diagnosis Diagnosis to resolution Initial hospital stay Total hospital stay
Aspirated (n = 7) 19 39 14 17
± ± ± ±
3 9 5 6
patient. During groin exploration a pulmonary embolus developed in one patient with unsuspected deep vein thrombosis after calf and thigh massage to promote vital dye transit. The interval from fistula diagnosis to resolution was more than two and one-half times longer in the nonoperative group than in the operative group (p = <0.01) (TableIV).Boththenumberofdaysof the initial hospital stay and the total hospital days were significantly greater in the nonoperative group. Wound infections occurred in five patients with a lymphocutaneous fistula treated without operation. The most common organisms were Staphylococcus aureus and Pseudomonas aeruginosa. In one patient with an expanded polytetrafluoroethylene femoropopliteal graft, a wound infection and subsequent graft infection that required graft removal and eventual limb amputation developed. No wound infection or infectious graft complication developed in any of the patients with a lymphocutaneous fistula treated with operation. Seven of the 13 patients with a lymphocele underwent needle aspiration and application of pressure dressings. The mean time interval from the initial operation until resolution was no different in the seven patients who underwent aspiration than in the six who did not undergo aspiration (Table V). Two patients had operative intervention; both had been previously aspirated. One patient was cured, but the second experienced a recurrence that resolved 35 days after operation. A wound infection developed in three of 13 patients with a lymphocele. There was no difference between those undergoing aspiration and those observed. No vascular grafts became infected.
17 35 19 20
± ± ± ±
5 10 3 3
NS NS NS NS
DISCUSSION The first reports of GLC after vascular operations appeared in the 1940s and described lymphocutaneous fistulas after venous ligations. 5,6 Croft reported the first GLC after an arterial reconstructive procedure in a patient in whom a lymphocutaneous fistula developed after femorofemoral bypass. 7 Since then, case reports and small series of GLC have appeared sporadically. Although uncommon, such complications are a well-recognized problem on most busy vascular surgery services. The exact incidence has not been previously determined. Kwaan and colleagues3 reported an incidence of 1.8% for lymphocutaneous fistulas alone, a finding similar to our combined lymphocutaneous fistula/lymphocele incidence of 1.5%. Because of the retrospective nature of our study, however, this value may in fact be an underestimate of total GLC and is probably a better estimate of severe lymphatic complications. The cause of GLC is unknown but presumably is due to disruption of lymphatic channels or lymph nodes during dissection. 8 Risk factors for development of GLC described in the literature include poor operative technique, inguinal adenopathy, scar tissue associated with repeat groin operations, and extensive groin dissections. 1. 3,9 The highest incidence oflymphatic complications in this series occurred in patients undergoing isolated femoral artery reconstruction in a scarred groin. Such redo procedures included anastomotic aneurysm repairs, graft thrombectomies and revisions, and common femoral artery-deep femoral endarterectomies, all procedures that usually require extensive groin dissection. Patients undergoing abdominal
862
Tyndall et at.
aortic aneurysm repair with ABF grafting after a prior groin operation also had a high incidence of GLC. One might theorize that aortic reconstruction with disruption of proximal lymphatic channels could increase the pressure in the groin lymphatics and increase the likelihood of a lymphatic leak. 10 Similarly, the low incidence of GLC after femoropopliteal-tibial bypass procedures even in scarred groins may result from disruption of lymphatics lower in the leg, which might lower the pressure in groin lymphatic channels. l l In this study lymphocutaneous fistulas were frequendy diagnosed during hospitalization, whereas lymphoceles were recognized later, usually during the first postoperative clinic visit. In Kwaan's3 series of lymphocutaneous fistulas, leaks were diagnosed within 3 days of operation, substantially earlier than in this series. The longer interval required to diagnose a lymphocutaneous fistula in this series assured exclusion of the more frequendy encountered minor lymphatic leaks and serous drainage that usually resolve without treatment. Optimal treatment of GLC has been controversial. Measures described for management of a lymphocutaneous fistula have included implantation of the damaged lymphatic channel into an adjacent vein, marsupialization and open packing, application of thrombotic agents, tight closure of the overlying tissues with pressure dressing, bed rest with leg elevation, observation, and ligation of the leaking lymphatic.l-4,12,13 Kwaan and colleagues3 were the first to document the improved outcome of lymphocutaneous fistulas treated with early reoperation. Despite these good results many surgeons still advise conservative treatment with bedrest, limb elevation, prophylactic antibiotics, and local wound care with pressure dressings. 1,2 The results of this study clearly argue against such a conservative approach and strongly support operative therapy. Nine of 10 patients treated with wound exploration had successful control of the leak site with significandy reduced hospital stays and no wound complications. Treatment of lymphoceles is also controversial and includes immobilization, repetitive aspirations, pressure dressings, percutaneous catheter drainage, radiation therapy, sclerosis, and operative excision of the cyst wall with ligation of the leak site. 14,15 Because lymphoceles are usually associated with a closed wound the risk of infectious complications seems less. Nevertheless, three patients with a lymphocele in this study developed superficial wound infections, although all three infections presented prior to recognition of the lymphocele. No lymphocele became
JOURNAL OF VASCULAR SURGERY May 1994
infected. The association between lymphoceles and wound infection is therefore unclear. Most patients with lymphoceles in this study were treated conservatively. Although recommended by some authorities, lymphocele aspiration followed by a pressure dressing did not hasten resolution. 16,17 On the other hand, aspiration did not increase the risk of infection and therefore may be useful for diagnostic purposes or when the lymphocele is large and symptomatic. Because of limited experience with reexploration, most surgeons favor conservative treatment except for large, progressive, and persistent lymphoceles. An argument against early operation for GLC is the fear of exposing the graft and subsequent graft infection. In this study all lymph leaks identified during exploration were found in the superficial subcutaneous layer, and exposure of the femoral vessels and grafts was not required. The superficial cutaneous lymphatics, which drain the skin and superficial fascia into the inferomedial and inferolateral superficial inguinal nodes, were the most commonly involved lymphatic chains. The deep lymphatics of the leg following along the branches of the femoral artery were not involved. Consequendy the risk of infecting a graft with wound exploration seems low. Although suggested by some, the use of vital dyes such as Evans blue or isosulfan blue to localize leaking lymphatics was not helpful in our experience. 3,4 Repeated massage of the thigh has been recommended as a useful technique for locating the leak site3; however, as described above in the patient with a pulmonary embolus, this technique is not without risk. Groin lymph leaks continue to be an uncommon but well-recognized complication after vascular surgical procedures. Once recognized, a persistent lymphocutaneous fistula should be treated by prompt reoperation with identification and control of the leak site. Treatment for a lymphocele should be individualized. Unless they are large, progressive, symptomatic, and persistent, most lymphoceles can be treated satisfactorily with conservative measures. Aspiration followed by a pressure dressing may occasionally be successful before considering operative exploration. REFERENCES 1. Wolfe JHN. Treatment oflympheelema. In: Rutherford RB, ed. Vascular surgery. 3rd eel. Philadelphia: WB Saunders; 1989:1668-78. 2. Johnston KW. Nonvascular complications of vascular surgery. In: Rutherfurd RB, ed. Vascular surgery. 3rd ed. Philadelphia: WB Saunders; 1989:536-40.
JOURNAL OF VASCULAR SURGERY Volume 19, Number 5
3. Kwaan JIIM, Bernstein JM, Connolly }E. Management of lymph fistula in the groin after arterial reconstruction. Arch Surg 1979;114:1416-8. 4. Weaver FA, Yellin AE. Management of postoperative lymphatic leaks by use ofisosulphan blue [Letter]. J VAse SURG 1991;14:566-7. 5. Harkins HN, Schug R. The surgical management of varicose veins: importance of individualization in the choice of procedure. Surgery 1942;11:402-21. 6. Fishback FC. Lymph leakage (lymphorrhea): a complication of saphenous vein ligation, with suggestions for treatment. Surgery 1947;22:834-6. 7. Croft RJ. Lymphatic fistula: a complication of arterial surgery [Letter]. Br Med J 1978;2:205. 8.. Baronofsky ID. Technique of inguinal node dissection. Surgery 1948;24:555-67. 9. Shah DM. In discussion: Kwaan JIIM, Bernstein JM, Conolly JE. Management of lymph fistula in the groin after arterial reconstruction. Arch Surg 1979;114:1418. 10. Jensen SR, Voegeli DR, McDermott JC, et al. Lymphatic disruption following abdominal aortic surgery. Cardiovasc Intervent RadiolI986;9:199-201. 11. Porter JM, Lindell TD, Lakin PC. Leg edema following
DISCUSSION Dr. Peter Gloviczki (Rochester, Minn.). As Dr. Tyndall pointed out, injury to the lymphatic system during exposure of the femoral vessels is almost unavoidable. Because of the rich lymphatic network of the femoral triangle, lymphatic complications after vascular reconstructions occur most frequently at the groin. The incidence of this in most studies similar to this experience is about 1%. It is higher in previously operated groins and higher if more extensive groin exploration is performed. I agree with most of the conclusions of the authors, most importantly with the aggressive management of lymphatic fistulas. The data that they presented, similar to several other reports in the literature, confirm that infection is more frequent if lymphatic fistulas are treated conservatively. Our approach at the Mayo Clinic is very similar. If local wound care, bed rest, and leg elevation are not effective to stop lymphatic fistula within a few days, surgical ligation is performed. In contrast to the authors' experience, however, we found that the injection of isosulfan blue dye into the foot is helpful to identify the site of the lymphatic leak. My question concerns the use of imaging techniques in the diagnosis oflymphatic complications. When would you recommend the use of ultrasonography, lymphoscintigraphy, or computed tomography? In our experience, if there is any doubt of an early or late lymphatic complication, lymphoscintigraphy confirms the diagnosis very accurately. Ultrasonography can be helpful to distinguish solid dense mass such as a hematoma from a cystic lymphocele. In patients diagnosed with late lymphocele or lymphatic fistula, computed tomography should also be performed,
Tyndall et al.
12.
13.
14.
15.
16.
17.
863
femoropopliteal autogenous vein bypass. Arch Surg 1972; 105:883-8. Judd ES, Nix JT. Spontaneous and traumatic lymph fistulas: data on 40 cases. Surg Clin North Am 1949;29:103547. Smellie GD, Wallace JR. Lymph fistulas and lymphocysts after peripheral vascular surgery. J R Coil Surg Edinb 1981;26: 78-81. VanSonnenberg E, Wittich GR, Casola G, et al. Lymphoceles: imaging characteristics and percutaneous management. Radiology 1986;161:593-6. Pope AJ, Ormiston MC, Bogod DG. Sclerotherapy in the treatment of a recurrent lymphocele. Postgrad Med J 1982; 58:573-4. Sethi GK, Scott SM, Takaro T. Persistent lymphatic fistula unusual complication of femoro-femoral arterial bypass. J Cardiovasc Surg 1978;19:155-9. DePalma RG. Disorders of the lymphatic system. In: Sabiston DC, ed. Textbook of surgery. 14th ed. Philadelphia: WB Saunders; 1991;1479-89.
Submitted Sept. 16, 1993; accepted Dec. 13, 1993.
not only to exclude graft infection, but also to diagnose a concomitant retroperitoneal lymphatic injury, because a retroperitoneal lymphocele or chylous ascites may occasionally present with lymphatic fistula at the groin. Finally, I agree with the authors that smalllymphoceles can be managed safely. In patients with enlarging or symptomatic lymphoceles, or those with a lymphocele immediately next to a prosthetic graft, however, we advocate early operation to reduce the risk of graft infection. Dr. Steve H. Tyndall. We believe that lymphocutaneous fistulas can be diagnosed clinically by the presence of a persistent clear drainage from the wound, and further diagnostic measures are not needed. Treatment oflymphocutaneous fistulas by reoperation without prior attempts at localization has been very successful in our hands. We have not used lymphoscintigraphy or magnetic resonance imaging for diagnosis, so I am unable to comment on its utility. Lymphoceles usually can be diagnosed clinically, but the use of color-flow duplex ultrasonography can be useful to distinguish a lymphocele from a hematoma or pseudoaneurysm. Diagnostic aspiration is helpful to differentiate infection from hematoma or seroma. The extensive transplant experience with the use of computed tomography scanning for abdominal lymphocele has shown it to be effective in diagnosis, determining the extent of the lymphocele and its anatomic relationships. Dr. James DeBord (Peoria, Ill.). You noted that you had significantly less groin lymphatic complications in your
JOURNAL OF VASCULAR SURGERY May 1994
864 Tyndall et at.
distal bypasses than in isolated groin explorations or proximal reconstructions. Do you believe this is because your lymphatic leak is occurring more distally in those patients and the lymph never gets to the groin? Did you have some distal lymphatic complications in those patients, and how do you manage them? Dr. Tyndall. None of the patients with groin lymphatic complications had a lymph leak lower in the leg, and we did not include patients with lymph leaks from incisions in other locations. We believe the higher incidence ofGLC
that occurs with isolated groin procedures and aortic aneurysm operations is related to relative groin lymphatic hypertension. With proximal disruption and ligation of lymphatics, the groin pressure increases and thus is more likely to develop a leak from a transected lymph vessel. On the other hand, patients with femoropopliteal or tibial bypass have their lymphatics disrupted lower in the leg and have a lower lymphatic pressure in the groin and a lower chance of a leak.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the JOURNAL OF VASCULAR SURGERY for 1994 are available to subscribers only. They may be purchased from the publisher at a cost of $76.00 for domestic, $97.32 for Canadian, and $92.00 for international subscribers for Vol. 19 (January to June) and Vol. 20 (July to December). Price includes shipping charges. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Subscription Services, Mosby-Year Book, Inc., 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA. In the United States call toll free (800)325-4177, ext. 4351. In Missouri or foreign countries call (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.
Groin lymphatic complications (GLC) are a rare but potentially serious problem after femoral arterial reconstruction. The risk of wound infection and subsequent graft infection, particularly of a synthetic vascular prosthesis, is a major concern. The cause of a lymphocutaneous fistula or lymphocele is presumed to be lymphatic disruption during groin dissection, but to date scant data are available to define risk factors. In addition controversy continues regarding optimal treatment, with some authorities favoring From the Department of Surgery, Division of Vascular Surgery, Henry Ford Hospital, Detroit. Presented at the Seventeenth Annual Meeting of the Midwestern Vascular Surgical Society, Chicago, Ill., Sept. 10-11, 1993. Reprint requests: Alexander D. Shepard, MD, Henry Ford Hospital, 2799 W. Grand Blvd., Detroit, MI 48202. Copyright © 1994 by The Society for Vascular Surgery and International Sociery for Cardiovascular Surgery, North American Chapter. 0741-5214/94/$3.00 + 0 24/6/53674
858
nonoperative therapyl,2 and others advocating early reoperation and direct control of the leak. 3,4 This review was undertaken in an effort to clarify these questions. MATERIAL AND METHODS The database of the Vascular Surgery Registry of the Henry Ford Hospital from 1977 through 1992 was retrospectively reviewed, and 41 patients with groin lymphatic complications after femoral arterial operations were identified. Data collected included patient demographics, operative indications, operation performed, presence of previous groin procedures, groin lymphatic complication treatment method, postoperative wound complications, and length of hospitalization and treatment. After arterial reconstruction all groin wounds were closed in three layers with interrupted absorbable sutures by either a senior resident or a staff
JOURNAL OF VASCULAR SURGERY Volume 19, Number 5
surgeon. Skin closure was based on surgeon preference and included staples, interrupted nylon stitches, or a running subcuticular absorbable suture. Lymphocutaneous fistulas were diagnosed by the appearance of persistent and near continuous clear drainage from a localized area of the groin wound most frequently near the distal end. Conservative treatment consisted of strict bed rest, limb elevation, prophylactic intravenous antibiotics, and aggressive local wound care, which included attempts at skin sterilization with povidone-iodine (Betadine) solution and pressure dressings. Operative therapy involved limited groin wound exploration. After removal of skin sutures and the first layer of subcutaneous stitches at the fistula site, draining fluid was tracked into the subcutaneous tissues, and an attempt was made to identify a leak site. Various vital dyes were injected into the first web space of the foot with subsequent calf/thigh compression in an attempt to promote operative identification. Presumed leak sites were suture ligated with absorbable suture. Exploration of the entire groin wound and removal of deep sutures closing tissue over the femoral artery or graft was avoided. After control of the presumed leaking lymphatic, the wound was meticulously closed. The choice of early or late wound exploration versus conservative management of the groin lymphatic complication was determined by individual staff surgeons. Lymphoceles were diagnosed by the presence of a persistent subcutaneous fluid collection in the absence of hematoma or wound abscess. These were either treated conservatively (bed rest, limited ambulation, or needle aspiration and pressure dressings) or by operative exploration, with excision of the lymphocele wall, attempted identification and control of the leak site, and meticulous wound closure. Data were statistically analyzed with the Student t test and Fisher's exact test. Data are expressed as the mean ± standard error. RESULTS Over the last 15 years a total of 2679 arterial reconstructive procedures requiring groin incisions were performed on the Vascular Surgery Service at the Henry Ford Hospital. Forty-one groin lymphatic complications consisting of 28 lymphocutaneous fistulas (68%) and 13 lymphoceles (32%) were recognized for an incidence of 1.5% per patient or 1.2% per groin incision. Of the 41 patients, 24 were men and 17 were women with an average age of 66 years (range 48 to 85 years). These sex and age distributions were not significantly different from
Tyndall et al.
859
those of patients undergoing femoral artery procedures in whom a GLC did not develop (Table I). Lymph complications were equally divided between the two groins, 21 on the right and 20 on the left. Mean follow-up after resolution of the groin lymphatic complication was 41 months (range 2 to 150 months). The percentage of redo groin operations, emergency procedures, and operations for tissue loss was no different for patients with and patients without lymphatic complications (Table I). Groin lymphatic problems occurred more frequently after aortobifemoral bypass (ABF) and isolated groin procedures (i.e., femoral anastomotic pseudoaneurysm repair, femoral endarterectomy, profundaplasty, graft thrombectomy, or embolectomy) than after infrainguinal bypasses (Table I). Although there was no overall difference in the incidence of groin lymphatic complications between patients with and without previous groin procedures, redo groin dissection did become a significant risk factor when analyzing specific operations. Patients undergoing an isolated femoral operation as a first groin procedure had a groin lymphatic complication incidence of 1.1 %, whereas those having an isolated femoral artery procedure after a previous groin operation had a 5.3% incidence (Table II). Similarly patients undergoing ABF grafting for abdominal aortic aneurysm repair had a groin lymphatic complication incidence of 0.7% in a nonoperated groin and 4.2% in a previously dissected groin (Table II). The lowest rate of groin lymphatic complications occurred in patients having femoropopliteal-tibial bypass grafting (0.5%). There was no difference in this group between patients with a previous femoral operation (0.3%) and those without one (0.7%) (Table II). When comparing patients with lymphocutaneous fistulas and lymphoceles, there was a significant difference in the number of days it took to diagnose the problem, 10 days for a lymphocutaneous fistula versus 18 days for a lymphocele (p = 0.02) (Table III). Most lymphocutaneous fistulas were diagnosed while patients were still hospitalized recovering from their reconstructive procedures (71 % ), whereas many lymphoceles were not diagnosed until the first outpatient visit (54%). Once a groin lymphatic complication was diagnosed, it took significantly longer for a lymphocele to resolve (37 ± 6 days) than it did for a lymphocutaneous fistula (20 ± 3 days) (p = 0.02). Five of eight patients diagnosed with a lymphocutaneous fistula after discharge from the hospital were readmitted for treatment, but only one
860
JOURNAL OF VASCULAR SURGERY May 1994
Tyndall et ai.
Table I. GLC: Patient demographics GLC (n = 41) Average age (yrs) Female Previous groin incision Tissue loss Emergency operation Operation performed Aortobifemoral graft Isolated femoral operation Femorodistal bypass
No GLC (n
= 2638)
P Value
66 (29%) (44%) (15%) (7%)
64 905 (34%) 1020 (39%) 524 (20%) 232 (9%)
NS NS NS NS NS
19 (46%) 15 (37%) 7 (17%)
665 (25%) 537 (20%) 1436 (54%)
<0.01 0.02 <0.01
12 18 6 3
Table II. Procedure-related incidence of GLC: Initial groin operations versus redo groin operations Initial groin operation
Redo groin operation
Per patient
Per incision
Per patient
Per incision
Operation (n)
(%)
(%)
(%)
(%)
P Value
All Operations (2679) APB (684) Isolated femoral (552) Femorodistal bypass (1443) APB for AAA (245) APB for occlusive disease (439)
1.4 2.7 1.1 0.7 1.4 3.5
1.0 1.4 1.1 0.7 0.7 1.8
1.7 3.0 5.3 0.3 8.1 1.5
1.5 1.5 5.3 0.3 4.2 0.8
NS NS <0.01
NS <0.05
NS
AAA, Abdominal aortic aneurysm.
Table III. Lymphocutaneous fistulas versus lymphoceles Interpal (mean days)
LF (n = 28)
Lymphocele (n = 13)
P Value
Initial operation to diagnosis Diagnosis to resolution Initial hospital stay Total hospital stay Infectious complications
10 ± 1 20 ± 3 16 ± 3 22 ± 2 18%
18 ± 3 37 ± 6 20 ± 2 18 ± 3 23%
0.02 0.02
of seven patients with a lymphocele discovered after discharge was rehospitalized. The interval from the original operation to resolution was significantly longer for a lymphocele (55 ± 9 days) than a lymphocutaneous fistula (30 ± 3 days). Skin closures at the time of the initial procedures in the 28 patients in whom a lymphocutaneous fistula developed included stapling in 13, interrupted nylon suturing in 13, and subcuticular absorbable suturing in two. Of the 13 patients in whom a lymphocele developed, five wounds were closed with staples, and eight were closed with interrupted nylon. Data on the method of groin wound closure was not available for patients without lymphatic complications. In general, however, staples and nylon sutures are used more frequently for groin wounds on our service than subcuticular closure. Eighteen patients with a lymphocutaneous fistula were treated without operation, and 10 underwent
NS NS NS
operative repair. There was no difference in the interval between the initial operation and the recognition of the lymph leak in these two groups (Table IV). Operative ligation of the leak site was successful in nine of 10 patients. At the time of reoperation, seven of these 10 patients had a leak site identified and ligated. In the three patients in whom no definite leak site was identified, a transected lymph node was found, though in no situation was it believed to be the definite source of the lymph leak. In two of these three patients the lymph drainage ceased within 24 hours of operation. In the other patient, lymph . drainage recurred rapidly after operation and required 24 days of conservative treatment before resolution, the average time it took for resolution in the nonoperative group. Injection of vital dyes did not help in the operative identification of leak sites. All identified lymph leaks were in the subcutaneous tissue. No graft or femoral artery was exposed in any
JOURNAL OF VASCULAR SURGERY
Tyndall et at.
Volume 19, Number 5
861
Table IV. Lymphocutaneous fistulas: Operative versus nonoperative therapy Interval (days) Initial operation to diagnosis Diagnosis to resolution Initial hospital stay Total hospital stay Infectious complications Graft infections
Operati-Pe (n = 10) 10 9 15 17
± ± ± ± 0 0
1 3 5 2
Nonopemti-Pe (n = 18)
P Value
10 ± 2 24 ± 3 23 ± 3 25 ± 3 5 (18%) 1 (6%)
0.84 <0.01 0.02 0.04 0.13 0.99
Nonaspirated (n = 6)
P Value
Table V. Lymphocele: Aspiration versus nonaspiration therapy Interval (mean days) Initial operation to diagnosis Diagnosis to resolution Initial hospital stay Total hospital stay
Aspirated (n = 7) 19 39 14 17
± ± ± ±
3 9 5 6
patient. During groin exploration a pulmonary embolus developed in one patient with unsuspected deep vein thrombosis after calf and thigh massage to promote vital dye transit. The interval from fistula diagnosis to resolution was more than two and one-half times longer in the nonoperative group than in the operative group (p = <0.01) (TableIV).Boththenumberofdaysof the initial hospital stay and the total hospital days were significantly greater in the nonoperative group. Wound infections occurred in five patients with a lymphocutaneous fistula treated without operation. The most common organisms were Staphylococcus aureus and Pseudomonas aeruginosa. In one patient with an expanded polytetrafluoroethylene femoropopliteal graft, a wound infection and subsequent graft infection that required graft removal and eventual limb amputation developed. No wound infection or infectious graft complication developed in any of the patients with a lymphocutaneous fistula treated with operation. Seven of the 13 patients with a lymphocele underwent needle aspiration and application of pressure dressings. The mean time interval from the initial operation until resolution was no different in the seven patients who underwent aspiration than in the six who did not undergo aspiration (Table V). Two patients had operative intervention; both had been previously aspirated. One patient was cured, but the second experienced a recurrence that resolved 35 days after operation. A wound infection developed in three of 13 patients with a lymphocele. There was no difference between those undergoing aspiration and those observed. No vascular grafts became infected.
17 35 19 20
± ± ± ±
5 10 3 3
NS NS NS NS
DISCUSSION The first reports of GLC after vascular operations appeared in the 1940s and described lymphocutaneous fistulas after venous ligations. 5,6 Croft reported the first GLC after an arterial reconstructive procedure in a patient in whom a lymphocutaneous fistula developed after femorofemoral bypass. 7 Since then, case reports and small series of GLC have appeared sporadically. Although uncommon, such complications are a well-recognized problem on most busy vascular surgery services. The exact incidence has not been previously determined. Kwaan and colleagues3 reported an incidence of 1.8% for lymphocutaneous fistulas alone, a finding similar to our combined lymphocutaneous fistula/lymphocele incidence of 1.5%. Because of the retrospective nature of our study, however, this value may in fact be an underestimate of total GLC and is probably a better estimate of severe lymphatic complications. The cause of GLC is unknown but presumably is due to disruption of lymphatic channels or lymph nodes during dissection. 8 Risk factors for development of GLC described in the literature include poor operative technique, inguinal adenopathy, scar tissue associated with repeat groin operations, and extensive groin dissections. 1. 3,9 The highest incidence oflymphatic complications in this series occurred in patients undergoing isolated femoral artery reconstruction in a scarred groin. Such redo procedures included anastomotic aneurysm repairs, graft thrombectomies and revisions, and common femoral artery-deep femoral endarterectomies, all procedures that usually require extensive groin dissection. Patients undergoing abdominal
862
Tyndall et at.
aortic aneurysm repair with ABF grafting after a prior groin operation also had a high incidence of GLC. One might theorize that aortic reconstruction with disruption of proximal lymphatic channels could increase the pressure in the groin lymphatics and increase the likelihood of a lymphatic leak. 10 Similarly, the low incidence of GLC after femoropopliteal-tibial bypass procedures even in scarred groins may result from disruption of lymphatics lower in the leg, which might lower the pressure in groin lymphatic channels. l l In this study lymphocutaneous fistulas were frequendy diagnosed during hospitalization, whereas lymphoceles were recognized later, usually during the first postoperative clinic visit. In Kwaan's3 series of lymphocutaneous fistulas, leaks were diagnosed within 3 days of operation, substantially earlier than in this series. The longer interval required to diagnose a lymphocutaneous fistula in this series assured exclusion of the more frequendy encountered minor lymphatic leaks and serous drainage that usually resolve without treatment. Optimal treatment of GLC has been controversial. Measures described for management of a lymphocutaneous fistula have included implantation of the damaged lymphatic channel into an adjacent vein, marsupialization and open packing, application of thrombotic agents, tight closure of the overlying tissues with pressure dressing, bed rest with leg elevation, observation, and ligation of the leaking lymphatic.l-4,12,13 Kwaan and colleagues3 were the first to document the improved outcome of lymphocutaneous fistulas treated with early reoperation. Despite these good results many surgeons still advise conservative treatment with bedrest, limb elevation, prophylactic antibiotics, and local wound care with pressure dressings. 1,2 The results of this study clearly argue against such a conservative approach and strongly support operative therapy. Nine of 10 patients treated with wound exploration had successful control of the leak site with significandy reduced hospital stays and no wound complications. Treatment of lymphoceles is also controversial and includes immobilization, repetitive aspirations, pressure dressings, percutaneous catheter drainage, radiation therapy, sclerosis, and operative excision of the cyst wall with ligation of the leak site. 14,15 Because lymphoceles are usually associated with a closed wound the risk of infectious complications seems less. Nevertheless, three patients with a lymphocele in this study developed superficial wound infections, although all three infections presented prior to recognition of the lymphocele. No lymphocele became
JOURNAL OF VASCULAR SURGERY May 1994
infected. The association between lymphoceles and wound infection is therefore unclear. Most patients with lymphoceles in this study were treated conservatively. Although recommended by some authorities, lymphocele aspiration followed by a pressure dressing did not hasten resolution. 16,17 On the other hand, aspiration did not increase the risk of infection and therefore may be useful for diagnostic purposes or when the lymphocele is large and symptomatic. Because of limited experience with reexploration, most surgeons favor conservative treatment except for large, progressive, and persistent lymphoceles. An argument against early operation for GLC is the fear of exposing the graft and subsequent graft infection. In this study all lymph leaks identified during exploration were found in the superficial subcutaneous layer, and exposure of the femoral vessels and grafts was not required. The superficial cutaneous lymphatics, which drain the skin and superficial fascia into the inferomedial and inferolateral superficial inguinal nodes, were the most commonly involved lymphatic chains. The deep lymphatics of the leg following along the branches of the femoral artery were not involved. Consequendy the risk of infecting a graft with wound exploration seems low. Although suggested by some, the use of vital dyes such as Evans blue or isosulfan blue to localize leaking lymphatics was not helpful in our experience. 3,4 Repeated massage of the thigh has been recommended as a useful technique for locating the leak site3; however, as described above in the patient with a pulmonary embolus, this technique is not without risk. Groin lymph leaks continue to be an uncommon but well-recognized complication after vascular surgical procedures. Once recognized, a persistent lymphocutaneous fistula should be treated by prompt reoperation with identification and control of the leak site. Treatment for a lymphocele should be individualized. Unless they are large, progressive, symptomatic, and persistent, most lymphoceles can be treated satisfactorily with conservative measures. Aspiration followed by a pressure dressing may occasionally be successful before considering operative exploration. REFERENCES 1. Wolfe JHN. Treatment oflympheelema. In: Rutherford RB, ed. Vascular surgery. 3rd eel. Philadelphia: WB Saunders; 1989:1668-78. 2. Johnston KW. Nonvascular complications of vascular surgery. In: Rutherfurd RB, ed. Vascular surgery. 3rd ed. Philadelphia: WB Saunders; 1989:536-40.
JOURNAL OF VASCULAR SURGERY Volume 19, Number 5
3. Kwaan JIIM, Bernstein JM, Connolly }E. Management of lymph fistula in the groin after arterial reconstruction. Arch Surg 1979;114:1416-8. 4. Weaver FA, Yellin AE. Management of postoperative lymphatic leaks by use ofisosulphan blue [Letter]. J VAse SURG 1991;14:566-7. 5. Harkins HN, Schug R. The surgical management of varicose veins: importance of individualization in the choice of procedure. Surgery 1942;11:402-21. 6. Fishback FC. Lymph leakage (lymphorrhea): a complication of saphenous vein ligation, with suggestions for treatment. Surgery 1947;22:834-6. 7. Croft RJ. Lymphatic fistula: a complication of arterial surgery [Letter]. Br Med J 1978;2:205. 8.. Baronofsky ID. Technique of inguinal node dissection. Surgery 1948;24:555-67. 9. Shah DM. In discussion: Kwaan JIIM, Bernstein JM, Conolly JE. Management of lymph fistula in the groin after arterial reconstruction. Arch Surg 1979;114:1418. 10. Jensen SR, Voegeli DR, McDermott JC, et al. Lymphatic disruption following abdominal aortic surgery. Cardiovasc Intervent RadiolI986;9:199-201. 11. Porter JM, Lindell TD, Lakin PC. Leg edema following
DISCUSSION Dr. Peter Gloviczki (Rochester, Minn.). As Dr. Tyndall pointed out, injury to the lymphatic system during exposure of the femoral vessels is almost unavoidable. Because of the rich lymphatic network of the femoral triangle, lymphatic complications after vascular reconstructions occur most frequently at the groin. The incidence of this in most studies similar to this experience is about 1%. It is higher in previously operated groins and higher if more extensive groin exploration is performed. I agree with most of the conclusions of the authors, most importantly with the aggressive management of lymphatic fistulas. The data that they presented, similar to several other reports in the literature, confirm that infection is more frequent if lymphatic fistulas are treated conservatively. Our approach at the Mayo Clinic is very similar. If local wound care, bed rest, and leg elevation are not effective to stop lymphatic fistula within a few days, surgical ligation is performed. In contrast to the authors' experience, however, we found that the injection of isosulfan blue dye into the foot is helpful to identify the site of the lymphatic leak. My question concerns the use of imaging techniques in the diagnosis oflymphatic complications. When would you recommend the use of ultrasonography, lymphoscintigraphy, or computed tomography? In our experience, if there is any doubt of an early or late lymphatic complication, lymphoscintigraphy confirms the diagnosis very accurately. Ultrasonography can be helpful to distinguish solid dense mass such as a hematoma from a cystic lymphocele. In patients diagnosed with late lymphocele or lymphatic fistula, computed tomography should also be performed,
Tyndall et al.
12.
13.
14.
15.
16.
17.
863
femoropopliteal autogenous vein bypass. Arch Surg 1972; 105:883-8. Judd ES, Nix JT. Spontaneous and traumatic lymph fistulas: data on 40 cases. Surg Clin North Am 1949;29:103547. Smellie GD, Wallace JR. Lymph fistulas and lymphocysts after peripheral vascular surgery. J R Coil Surg Edinb 1981;26: 78-81. VanSonnenberg E, Wittich GR, Casola G, et al. Lymphoceles: imaging characteristics and percutaneous management. Radiology 1986;161:593-6. Pope AJ, Ormiston MC, Bogod DG. Sclerotherapy in the treatment of a recurrent lymphocele. Postgrad Med J 1982; 58:573-4. Sethi GK, Scott SM, Takaro T. Persistent lymphatic fistula unusual complication of femoro-femoral arterial bypass. J Cardiovasc Surg 1978;19:155-9. DePalma RG. Disorders of the lymphatic system. In: Sabiston DC, ed. Textbook of surgery. 14th ed. Philadelphia: WB Saunders; 1991;1479-89.
Submitted Sept. 16, 1993; accepted Dec. 13, 1993.
not only to exclude graft infection, but also to diagnose a concomitant retroperitoneal lymphatic injury, because a retroperitoneal lymphocele or chylous ascites may occasionally present with lymphatic fistula at the groin. Finally, I agree with the authors that smalllymphoceles can be managed safely. In patients with enlarging or symptomatic lymphoceles, or those with a lymphocele immediately next to a prosthetic graft, however, we advocate early operation to reduce the risk of graft infection. Dr. Steve H. Tyndall. We believe that lymphocutaneous fistulas can be diagnosed clinically by the presence of a persistent clear drainage from the wound, and further diagnostic measures are not needed. Treatment oflymphocutaneous fistulas by reoperation without prior attempts at localization has been very successful in our hands. We have not used lymphoscintigraphy or magnetic resonance imaging for diagnosis, so I am unable to comment on its utility. Lymphoceles usually can be diagnosed clinically, but the use of color-flow duplex ultrasonography can be useful to distinguish a lymphocele from a hematoma or pseudoaneurysm. Diagnostic aspiration is helpful to differentiate infection from hematoma or seroma. The extensive transplant experience with the use of computed tomography scanning for abdominal lymphocele has shown it to be effective in diagnosis, determining the extent of the lymphocele and its anatomic relationships. Dr. James DeBord (Peoria, Ill.). You noted that you had significantly less groin lymphatic complications in your
JOURNAL OF VASCULAR SURGERY May 1994
864 Tyndall et at.
distal bypasses than in isolated groin explorations or proximal reconstructions. Do you believe this is because your lymphatic leak is occurring more distally in those patients and the lymph never gets to the groin? Did you have some distal lymphatic complications in those patients, and how do you manage them? Dr. Tyndall. None of the patients with groin lymphatic complications had a lymph leak lower in the leg, and we did not include patients with lymph leaks from incisions in other locations. We believe the higher incidence ofGLC
that occurs with isolated groin procedures and aortic aneurysm operations is related to relative groin lymphatic hypertension. With proximal disruption and ligation of lymphatics, the groin pressure increases and thus is more likely to develop a leak from a transected lymph vessel. On the other hand, patients with femoropopliteal or tibial bypass have their lymphatics disrupted lower in the leg and have a lower lymphatic pressure in the groin and a lower chance of a leak.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the JOURNAL OF VASCULAR SURGERY for 1994 are available to subscribers only. They may be purchased from the publisher at a cost of $76.00 for domestic, $97.32 for Canadian, and $92.00 for international subscribers for Vol. 19 (January to June) and Vol. 20 (July to December). Price includes shipping charges. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Subscription Services, Mosby-Year Book, Inc., 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA. In the United States call toll free (800)325-4177, ext. 4351. In Missouri or foreign countries call (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.