- Email: [email protected]
Peritoneo-cystic shunt for malignant ascites
GYNECOLOGICONCOLOGY18, 402-407 (1984)
Peritoneo-cystic FREDERICK Department
B.
STEHMAN,
Shunt for Malignant Ascites’ M.D.,
AND CLARENCE
of Obstetrics and Gynecology, Indiana 926 West Michigan Street, Indianapolis,
E. EHRLICH, M.D.
University Medical Indiana 46223
Center,
Received June 8, 1983 Peritoneo-venous shunting has been used extensively in the treatment of benign ascites and, to a limited extent, in the palliative management of malignant ascites. Acceptance of this therapy for malignant ascites has been slow because of concern over intravascular dissemination of disease. Recently a patient with advanced drug-resistant ovarian carcinoma was treated with peritoneo-cystic shunt. This patient’s tumor had progressed on multiple chemotherapeutic agents. She continued to work 40 hr per week but her activity was limited by massive ascites. The Denver Shunt (Storz) was selected in preference to the strut-type shunt. The Denver Shunt has a miter valve which is less likely to become occluded by fibrinous and cellular debris, and manual compression of the pumping chamber allows flushing and control of flow. This patient’s shunt remained patent for 5 months, until her death, documented by urine cytology and cystoscopy. Initial control of ascites was only fair, probably due to the virtual absence of a pressure gradient between the peritoneal cavity and the bladder. Without a pressure gradient, spontaneous flow would be expected to be nil. Though feasible and well tolerated, this technique is probably not useful in the management of malignant ascites. If modifications of the device could be made to increase the manual flow rate, then this technique might be acceptable.
In spite of advances in local and systemic therapies, the control or palliation of malignant ascites continues to be a significant problem. When the underlying disease process can be controlled by surgery, radiation therapy, or chemotherapy, ascites is also controlled. However, local measures directed at the ascites itself have been less successful than local treatment of malignant pleural effusion. Medical management, fluid or sodium restriction and diuretics, has been of limited efficacy and tends to further deplete circulating blood volume. LeVeen popularized the use of peritoneo-venous shunting of ascites through a pressure-activated silicone-strut valve [I ,2]. This procedure has become a valuable adjunct to the management of cirrhotic ascites or ascites due to Budd-Chiari Syndrome. Acceptance has been slow in the treatment of malignant ascites, however, largely because of concern over the possibility of dissemination of disease. We have recently treated a patient with ovarian carcinoma and uncontrolled ascites by performing a peritoneo-cystic shunt using a pressure-activated miter valve (Denver Shunt Storz). While this procedure was partially successful, it did not address the fluid and protein loss associated with repeated paracenteses. ’ Presented at the Tenth Annual Meeting of the Western Association of Gynecologic Oncologists Santa Fe, N. Mex., May 21, 1982. 0090-8258184$1.50 Copyright All rights
0 1984 by Academic Press, Inc. of reproduction in any form reserved.
402
CASE REPORTS
403
CASE REPORT
The patient, a 46-year-old GiP, , originally presented to her personal physician with ascites, bilateral adnexal masses, and a partial colon obstruction. Bilateral salpingo-oophorectomy, supracervical hysterectomy, and omental biopsy were performed for a stage III grade 2 papillary serous ovarian adenocarcinoma. Combination chemotherapy consisting of cyclophosphamide, hexamethylmelamine, doxorubicin, and cisplatin was initiated, but hexamethylmelamine was deleted because of marrow suppression. The patient achieved complete clinical response and, after six cycles of therapy, underwent reexploration (‘second-look” laparotomy). Persistent small volume disease was identified and five more cycles of chemotherapy were given. Another reexploration (“third look”) was carried out and disease was still present. At the time of referral to Indiana University Medical Center, the patient was asymptomatic and free of clinically measurable disease. She was treated with oral cyclophosphamide maintenance until progression of disease. Hexamethylmelamine was reinstituted. Tumor continued to progress and therapy with the investigational drug N-(phosphonacetyl)-L-aspartate (PALA) was initiated. The patient continued to work 40 hr per week but her activity was limited by ascites. The pelvic tumor and the ascites progressed in spite of therapy with PALA. Ascites reaccumulated rapidly after paracentesis. It was thought that the ascites was the aspect of the patient’s disease causing the greatest symptoms. We elected to perform a peritoneo-cystic shunt for palliation. Postoperatively, shunt patency could be documented by the presence of malignant cells in the urine. After vigorous compression of the shunt chamber, the urine could be seen to be cloudy and blood-tinged. Initially there was moderate control of ascites, but within a month repeat paracentesis was required. Pedal edema and peau d’orange change of the lower abdomen were first noted about 1 month postoperatively. The patient continued to require narcotic analgesics. Tumor-directed therapy was continued with Sfluorouracil (5-FU) and, later, tamoxifen. Repeated paracenteses were required as tumor continued to progress. Four months after the placement of the shunt, the pump chamber could not be palpated well due to lower abdominal edema. Cystoscopy confirmed patency of the shunt and absence of gross tumor in the bladder. The patient expired 5 months after placement of the shunt. A limited postmortem exam showed no tumor in the bladder or in the subcutaneous tunnel, The shunt functioned well after removal and was free of fibrinous sheath, clot, or cellular obstruction. OPERATIVE TECHNIQUE
General anesthesia was induced with the patient in the supine position. A longitudinal incision was made over the right anterior superior iliac spine. A small transverse incision was made over the right flank approximately 8 cm higher. A subcutaneous tunnel was made between these two incisions and the peritoneal limb of tubing was passed through the tunnel into the peritoneal cavity and secured with a purse-string suture. The pump chamber was secured to the
404
STEHMAN
AND
EHRLICH
periosteum of the iliac spine. A transverse incision was made above the symphysis over the dome of the bladder. A flap was developed similar to a Boari flap for ureteric reimplantation or a Janeway tube gastrostomy (Figs. 1, 2). The venous limb was secured to this flap and the incisions were closed. With pressure on the abdomen, bloody ascitic fluid could easily be seen to flow from the Foley catheter. DISCUSSION
The inadequacy of medical management of ascites is attested to by the numerous attempts at surgical therapy. Smith first proposed the use of a valved shunt in 1962 [3]. It was LeVeen who modified the valve and popularized peritoneovenous shunting. LeVeen’s shunt consists of a pressure-activated silicone-strut valve and tubing limbs to the peritoneal cavity and to the venous system. This valve has no pumping action and reduces the risk of forced embolus. A pressure gradient of 3 cm of water activates flow and there is virtually no reflux. Increased
FIG.
1. Bladder exposed by extraperitoneal approach with flap outlined (dashed line).
CASE
FIG.
2.
REPORTS
405
Shunt tubing through subcutaneous tunnel ready to be secured to flap.
flow can be achieved by respiratory exercises that increase the pressure gradient between the thoracic and peritoneal cavities. The strut valve may stick open or closed if fibrin or cellular debris collects in the valve chamber. Then there is no option except to replace the entire shunt. The Denver Shunt, developed by Dr. Newkirk [4], uses a miter valve in a soft silicone pump chamber. A pressure gradient of approximately 1 cm of water activates flow. Again, there is virtually no reflux. The chamber can be “flushed” by applying manual pressure on the reservoir. Such pump compression can augment flow and dislodge any fibrinous or cellular debris that could render a valve incompetent. This type of shunt has an advantage in the treatment of cellular or bloody ascites. Peritoneo-venous shunts are not free of complications, however. Pulmonary edema, coagulopathy, thrombosis, embolus, and infection have been reported in addition to valve dysfunction or failure. In malignant ascites, dissemination of disease may occur. This complication is well recognized in ventriculo-atria1 shunting and has been reported in peritoneo-venous shunting.
406
STEHMAN
AND
EHRLICH
The reported experience with peritoneo-venous shunting for malignant ascites totals 172 patients (Table 1). Occasional patients have derived long-term benefit, but there appear to be a number of patients who have not benefited or whose conditions have been aggravated by the procedure. Autopsy documentation of metastasis through the shunt to the vena cava and into the pulmonary vasculature has been provided. This is a recognized complication of ventriculo-atria1 shunts for brain tumors. Since ovarian carcinoma tends to remain in the peritoneal cavity, metastasis could be of special concern in dealing with that group of patients. Compared to patients with acellular ascites, patients with cellular or bloody ascites do poorly. Loculated ascites and protein-rich ascites also appear to be associated with a lower success rate. Ascites in ovarian carcinoma is frequently cellular, bloody, loculated, and protein rich. The idea of draining ascites through the bladder rather than performing repeated paracenteses is not new. Mulvaney performed vesico-coelomic drainage for 11 patients, reporting his results in 1955 [16]. Eight patients had malignant ascites. He used a valveless catheter and documented function by instilling dextrose intraperitoneally and monitoring dextrose in the urine. He also noted albuminuria, malignant cells in the urine, and freedom from subsequent paracentesis. No tumor implants in the bladder were observed. TABLE
Author Hyde and Eiseman [.5] Pollock [6] Arnot and White [7] LeVeen et al. [2] Lund and Newkirk [4] Strauss et al. [8]
Number of patients 1 9 3 18 10 37
Number of ovarian cancers 0
2 1 not spec 5 13
Maat el al. 191
1
1
Raaf and Stroehlein [lo]
5
0
3
3
8
3
10
Lokich et al. [ll]
1
Kudsk et al. [12] Qazi and Savlov [13]
40
3 28
Gough [14] Cheung and Raaf [IS]
10 17
2 2
172
63
Comments Survived 8 weeks 4 early deaths Ovarian failed 1 metastasis in vena cava 1 survived 7 months 2 failures-loculated ascites; 4 no function; 4 function <3 months Patient died at 14 days with multiple tumor emboli in lungs These referred patients had pulmonary metastases and tumor in the subcutaneous tunnel; the authors felt ovarian cancer to be a contraindication. 2 patients with tumor in lung at autopsy; two-thirds ovarian patients did poorly Two-thirds ovarian patients failed 12 patients derived no benefit 2 patients with caval metastases 1 survived 15 months 14% survive 6 months
CASE REPORTS
407
We believe that ours is the first case in which a valved shunt has been used in this fashion. While we have demonstrated the feasibility of the procedure and have documented patency for 5 months, the efficacy in terms of relieving ascites was less than satisfactory. Even though this kind of valve can be activated by a low-pressure gradient, little spontaneous flow occurred. Anatomically, the bladder is subject to intraabdominal pressure and virtually no gradient exists. Manual compression of the reservoir did not drain sufficient ascites to provide palliative relief. Perhaps a modification of the reservoir that would allow greater drainage by manual compression could be effected. In this case, progressive lower abdominal edema made it more difficult to identify and compress the reservoir. Even though some palliation of the ascites was achieved, this symptom was replaced by increasing disability due to pain. The disfigurement of massive ascites focuses our attention upon that symptom while we may underestimate other, equally serious, symptoms. Peritoneo-venous shunting, if successful, could avert the risks of tumor dissemination, pulmonary edema, and thromboembolism. It would not address the protein and fluid loss associated with repeated paracentesis. REFERENCES 1. LeVeen, H. H., Christoudias, G., Ip, M., Luft, R., Falk, G., and Grosberg, S. Peritoneo-venous shunting for ascites, Ann. Sup. 180, 580-591 (1974). 2. LeVeen, H. H., Wapnick, S., Grosberg, S., and Kinney, M. J. Further experience with peritoneovenous shunt for ascites, Ann. Surg. 184, 574-581 (1976). 3. Smith, A. N., Preshaw, R. M., and Bisset, W. H. The drainage of malignant ascites by a modification of the Spitz-Holter valve technique, J. R. Co/l. Surg. Edinburgh 7, 289-294 (1%2). 4. Lund, R. H., and Newkirk, J. B. Peritoneo-venous shunting system for surgical managment of ascites, Contemp. Surg. 14, 3145 (1979). 5. Hyde, G. L., and Eiseman, B. Peritoneal atria1 shunt for intractable ascites, Arch. Surg. 95, 369-373 (1%7). 6. Pollock, A. V. The treatment of resistant malignant ascites by insertion of a peritoneo-atria1 Holter valve. Brit. J. Surg. 62, 104-107 (1975). 7. Amot, R. S., and White, H. LeVeen shunts, Lancer 1, 505 (1978). 8. Strauss, A. K., Roseman, D. L., and Shapiro, T. M. Peritoneo-venous shunting in the management of ascites, Arch. Surg. 114, 48W91 (1979). 9. Maat, B., Oosterlee, J., Spaas, J. A. J., White, H., and Lammes, F. B. Dissemination of tumor cells with LeVeen shunt. Lancer 1, 988 (1979). 10. Raaf, J. H., and Stroehlein, J. R. Palliation of malignant ascites by the LeVeen peritoneo-venous shunt, Cancer 45, 1019-1024 (1980). 11. Lokich, J., Reinhold, R., Silverman, M., and Tullis, J. Complications of peritoneovenous shunts for malignant ascites, Cancer Treat. Rep. 64, 3W309 (1980). 12. Kudsk, K., Fabian, T. C., and Minton, J. P. LeVeen shunts in patients with intractable malignant ascites, J. Surg. Oncol. 13, 61-66 (1980). 13. Qazi, R., and Savlov, E. D. Peritoneovenous shunt for palliation of malignant ascites, Cancer 49, 600-602 (1982). 14. Gough, I. R. Peritoneo-venous shunts for malignant ascites, Amt. N.Z. J. Surg. 52,47-49 (1982). 15. Cheung, D. K., and Raaf, J. H. Selection of patients with malignant ascites for a peritoneovenous shunt, Cancer 50, 12w1209 (1982). 16. Mulvaney, D. Vesico-coelomic drainage for the ,relief of ascites, Lancer 2, 748-749 (1955).
Peritoneo-cystic FREDERICK Department
B.
STEHMAN,
Shunt for Malignant Ascites’ M.D.,
AND CLARENCE
of Obstetrics and Gynecology, Indiana 926 West Michigan Street, Indianapolis,
E. EHRLICH, M.D.
University Medical Indiana 46223
Center,
Received June 8, 1983 Peritoneo-venous shunting has been used extensively in the treatment of benign ascites and, to a limited extent, in the palliative management of malignant ascites. Acceptance of this therapy for malignant ascites has been slow because of concern over intravascular dissemination of disease. Recently a patient with advanced drug-resistant ovarian carcinoma was treated with peritoneo-cystic shunt. This patient’s tumor had progressed on multiple chemotherapeutic agents. She continued to work 40 hr per week but her activity was limited by massive ascites. The Denver Shunt (Storz) was selected in preference to the strut-type shunt. The Denver Shunt has a miter valve which is less likely to become occluded by fibrinous and cellular debris, and manual compression of the pumping chamber allows flushing and control of flow. This patient’s shunt remained patent for 5 months, until her death, documented by urine cytology and cystoscopy. Initial control of ascites was only fair, probably due to the virtual absence of a pressure gradient between the peritoneal cavity and the bladder. Without a pressure gradient, spontaneous flow would be expected to be nil. Though feasible and well tolerated, this technique is probably not useful in the management of malignant ascites. If modifications of the device could be made to increase the manual flow rate, then this technique might be acceptable.
In spite of advances in local and systemic therapies, the control or palliation of malignant ascites continues to be a significant problem. When the underlying disease process can be controlled by surgery, radiation therapy, or chemotherapy, ascites is also controlled. However, local measures directed at the ascites itself have been less successful than local treatment of malignant pleural effusion. Medical management, fluid or sodium restriction and diuretics, has been of limited efficacy and tends to further deplete circulating blood volume. LeVeen popularized the use of peritoneo-venous shunting of ascites through a pressure-activated silicone-strut valve [I ,2]. This procedure has become a valuable adjunct to the management of cirrhotic ascites or ascites due to Budd-Chiari Syndrome. Acceptance has been slow in the treatment of malignant ascites, however, largely because of concern over the possibility of dissemination of disease. We have recently treated a patient with ovarian carcinoma and uncontrolled ascites by performing a peritoneo-cystic shunt using a pressure-activated miter valve (Denver Shunt Storz). While this procedure was partially successful, it did not address the fluid and protein loss associated with repeated paracenteses. ’ Presented at the Tenth Annual Meeting of the Western Association of Gynecologic Oncologists Santa Fe, N. Mex., May 21, 1982. 0090-8258184$1.50 Copyright All rights
0 1984 by Academic Press, Inc. of reproduction in any form reserved.
402
CASE REPORTS
403
CASE REPORT
The patient, a 46-year-old GiP, , originally presented to her personal physician with ascites, bilateral adnexal masses, and a partial colon obstruction. Bilateral salpingo-oophorectomy, supracervical hysterectomy, and omental biopsy were performed for a stage III grade 2 papillary serous ovarian adenocarcinoma. Combination chemotherapy consisting of cyclophosphamide, hexamethylmelamine, doxorubicin, and cisplatin was initiated, but hexamethylmelamine was deleted because of marrow suppression. The patient achieved complete clinical response and, after six cycles of therapy, underwent reexploration (‘second-look” laparotomy). Persistent small volume disease was identified and five more cycles of chemotherapy were given. Another reexploration (“third look”) was carried out and disease was still present. At the time of referral to Indiana University Medical Center, the patient was asymptomatic and free of clinically measurable disease. She was treated with oral cyclophosphamide maintenance until progression of disease. Hexamethylmelamine was reinstituted. Tumor continued to progress and therapy with the investigational drug N-(phosphonacetyl)-L-aspartate (PALA) was initiated. The patient continued to work 40 hr per week but her activity was limited by ascites. The pelvic tumor and the ascites progressed in spite of therapy with PALA. Ascites reaccumulated rapidly after paracentesis. It was thought that the ascites was the aspect of the patient’s disease causing the greatest symptoms. We elected to perform a peritoneo-cystic shunt for palliation. Postoperatively, shunt patency could be documented by the presence of malignant cells in the urine. After vigorous compression of the shunt chamber, the urine could be seen to be cloudy and blood-tinged. Initially there was moderate control of ascites, but within a month repeat paracentesis was required. Pedal edema and peau d’orange change of the lower abdomen were first noted about 1 month postoperatively. The patient continued to require narcotic analgesics. Tumor-directed therapy was continued with Sfluorouracil (5-FU) and, later, tamoxifen. Repeated paracenteses were required as tumor continued to progress. Four months after the placement of the shunt, the pump chamber could not be palpated well due to lower abdominal edema. Cystoscopy confirmed patency of the shunt and absence of gross tumor in the bladder. The patient expired 5 months after placement of the shunt. A limited postmortem exam showed no tumor in the bladder or in the subcutaneous tunnel, The shunt functioned well after removal and was free of fibrinous sheath, clot, or cellular obstruction. OPERATIVE TECHNIQUE
General anesthesia was induced with the patient in the supine position. A longitudinal incision was made over the right anterior superior iliac spine. A small transverse incision was made over the right flank approximately 8 cm higher. A subcutaneous tunnel was made between these two incisions and the peritoneal limb of tubing was passed through the tunnel into the peritoneal cavity and secured with a purse-string suture. The pump chamber was secured to the
404
STEHMAN
AND
EHRLICH
periosteum of the iliac spine. A transverse incision was made above the symphysis over the dome of the bladder. A flap was developed similar to a Boari flap for ureteric reimplantation or a Janeway tube gastrostomy (Figs. 1, 2). The venous limb was secured to this flap and the incisions were closed. With pressure on the abdomen, bloody ascitic fluid could easily be seen to flow from the Foley catheter. DISCUSSION
The inadequacy of medical management of ascites is attested to by the numerous attempts at surgical therapy. Smith first proposed the use of a valved shunt in 1962 [3]. It was LeVeen who modified the valve and popularized peritoneovenous shunting. LeVeen’s shunt consists of a pressure-activated silicone-strut valve and tubing limbs to the peritoneal cavity and to the venous system. This valve has no pumping action and reduces the risk of forced embolus. A pressure gradient of 3 cm of water activates flow and there is virtually no reflux. Increased
FIG.
1. Bladder exposed by extraperitoneal approach with flap outlined (dashed line).
CASE
FIG.
2.
REPORTS
405
Shunt tubing through subcutaneous tunnel ready to be secured to flap.
flow can be achieved by respiratory exercises that increase the pressure gradient between the thoracic and peritoneal cavities. The strut valve may stick open or closed if fibrin or cellular debris collects in the valve chamber. Then there is no option except to replace the entire shunt. The Denver Shunt, developed by Dr. Newkirk [4], uses a miter valve in a soft silicone pump chamber. A pressure gradient of approximately 1 cm of water activates flow. Again, there is virtually no reflux. The chamber can be “flushed” by applying manual pressure on the reservoir. Such pump compression can augment flow and dislodge any fibrinous or cellular debris that could render a valve incompetent. This type of shunt has an advantage in the treatment of cellular or bloody ascites. Peritoneo-venous shunts are not free of complications, however. Pulmonary edema, coagulopathy, thrombosis, embolus, and infection have been reported in addition to valve dysfunction or failure. In malignant ascites, dissemination of disease may occur. This complication is well recognized in ventriculo-atria1 shunting and has been reported in peritoneo-venous shunting.
406
STEHMAN
AND
EHRLICH
The reported experience with peritoneo-venous shunting for malignant ascites totals 172 patients (Table 1). Occasional patients have derived long-term benefit, but there appear to be a number of patients who have not benefited or whose conditions have been aggravated by the procedure. Autopsy documentation of metastasis through the shunt to the vena cava and into the pulmonary vasculature has been provided. This is a recognized complication of ventriculo-atria1 shunts for brain tumors. Since ovarian carcinoma tends to remain in the peritoneal cavity, metastasis could be of special concern in dealing with that group of patients. Compared to patients with acellular ascites, patients with cellular or bloody ascites do poorly. Loculated ascites and protein-rich ascites also appear to be associated with a lower success rate. Ascites in ovarian carcinoma is frequently cellular, bloody, loculated, and protein rich. The idea of draining ascites through the bladder rather than performing repeated paracenteses is not new. Mulvaney performed vesico-coelomic drainage for 11 patients, reporting his results in 1955 [16]. Eight patients had malignant ascites. He used a valveless catheter and documented function by instilling dextrose intraperitoneally and monitoring dextrose in the urine. He also noted albuminuria, malignant cells in the urine, and freedom from subsequent paracentesis. No tumor implants in the bladder were observed. TABLE
Author Hyde and Eiseman [.5] Pollock [6] Arnot and White [7] LeVeen et al. [2] Lund and Newkirk [4] Strauss et al. [8]
Number of patients 1 9 3 18 10 37
Number of ovarian cancers 0
2 1 not spec 5 13
Maat el al. 191
1
1
Raaf and Stroehlein [lo]
5
0
3
3
8
3
10
Lokich et al. [ll]
1
Kudsk et al. [12] Qazi and Savlov [13]
40
3 28
Gough [14] Cheung and Raaf [IS]
10 17
2 2
172
63
Comments Survived 8 weeks 4 early deaths Ovarian failed 1 metastasis in vena cava 1 survived 7 months 2 failures-loculated ascites; 4 no function; 4 function <3 months Patient died at 14 days with multiple tumor emboli in lungs These referred patients had pulmonary metastases and tumor in the subcutaneous tunnel; the authors felt ovarian cancer to be a contraindication. 2 patients with tumor in lung at autopsy; two-thirds ovarian patients did poorly Two-thirds ovarian patients failed 12 patients derived no benefit 2 patients with caval metastases 1 survived 15 months 14% survive 6 months
CASE REPORTS
407
We believe that ours is the first case in which a valved shunt has been used in this fashion. While we have demonstrated the feasibility of the procedure and have documented patency for 5 months, the efficacy in terms of relieving ascites was less than satisfactory. Even though this kind of valve can be activated by a low-pressure gradient, little spontaneous flow occurred. Anatomically, the bladder is subject to intraabdominal pressure and virtually no gradient exists. Manual compression of the reservoir did not drain sufficient ascites to provide palliative relief. Perhaps a modification of the reservoir that would allow greater drainage by manual compression could be effected. In this case, progressive lower abdominal edema made it more difficult to identify and compress the reservoir. Even though some palliation of the ascites was achieved, this symptom was replaced by increasing disability due to pain. The disfigurement of massive ascites focuses our attention upon that symptom while we may underestimate other, equally serious, symptoms. Peritoneo-venous shunting, if successful, could avert the risks of tumor dissemination, pulmonary edema, and thromboembolism. It would not address the protein and fluid loss associated with repeated paracentesis. REFERENCES 1. LeVeen, H. H., Christoudias, G., Ip, M., Luft, R., Falk, G., and Grosberg, S. Peritoneo-venous shunting for ascites, Ann. Sup. 180, 580-591 (1974). 2. LeVeen, H. H., Wapnick, S., Grosberg, S., and Kinney, M. J. Further experience with peritoneovenous shunt for ascites, Ann. Surg. 184, 574-581 (1976). 3. Smith, A. N., Preshaw, R. M., and Bisset, W. H. The drainage of malignant ascites by a modification of the Spitz-Holter valve technique, J. R. Co/l. Surg. Edinburgh 7, 289-294 (1%2). 4. Lund, R. H., and Newkirk, J. B. Peritoneo-venous shunting system for surgical managment of ascites, Contemp. Surg. 14, 3145 (1979). 5. Hyde, G. L., and Eiseman, B. Peritoneal atria1 shunt for intractable ascites, Arch. Surg. 95, 369-373 (1%7). 6. Pollock, A. V. The treatment of resistant malignant ascites by insertion of a peritoneo-atria1 Holter valve. Brit. J. Surg. 62, 104-107 (1975). 7. Amot, R. S., and White, H. LeVeen shunts, Lancer 1, 505 (1978). 8. Strauss, A. K., Roseman, D. L., and Shapiro, T. M. Peritoneo-venous shunting in the management of ascites, Arch. Surg. 114, 48W91 (1979). 9. Maat, B., Oosterlee, J., Spaas, J. A. J., White, H., and Lammes, F. B. Dissemination of tumor cells with LeVeen shunt. Lancer 1, 988 (1979). 10. Raaf, J. H., and Stroehlein, J. R. Palliation of malignant ascites by the LeVeen peritoneo-venous shunt, Cancer 45, 1019-1024 (1980). 11. Lokich, J., Reinhold, R., Silverman, M., and Tullis, J. Complications of peritoneovenous shunts for malignant ascites, Cancer Treat. Rep. 64, 3W309 (1980). 12. Kudsk, K., Fabian, T. C., and Minton, J. P. LeVeen shunts in patients with intractable malignant ascites, J. Surg. Oncol. 13, 61-66 (1980). 13. Qazi, R., and Savlov, E. D. Peritoneovenous shunt for palliation of malignant ascites, Cancer 49, 600-602 (1982). 14. Gough, I. R. Peritoneo-venous shunts for malignant ascites, Amt. N.Z. J. Surg. 52,47-49 (1982). 15. Cheung, D. K., and Raaf, J. H. Selection of patients with malignant ascites for a peritoneovenous shunt, Cancer 50, 12w1209 (1982). 16. Mulvaney, D. Vesico-coelomic drainage for the ,relief of ascites, Lancer 2, 748-749 (1955).