Anda di halaman 1dari 8

GYNECOLOGIC ONCOLOGY 64, 80–87 (1997) ARTICLE NO. GO964529

A Randomized Trial of Paracentesis plus Intraperitoneal Tumor Necrosis Factor-a versus Paracentesis Alone in Patients with Symptomatic Ascites from Recurrent Ovarian Carcinoma*

H. W. HIRTE, 1 D. MILLER, 2 K. TONKIN , 3 B. FINDLAY , 4 V. CAPSTICK , 5 J. MURPHY , 6 R. BUCKMAN , J. CARMICHAEL, 8 M. LEVINE, 9 AND W. HILL 10

7

1 Ontario Cancer Treatment and Research Foundation, Hamilton Regional Cancer Centre, Hamilton, Ontario; 2 Cancer Control Agency of British Columbia, Vancouver, British Columbia; 3 Ontario Cancer Treatment and Research Foundation, London Regional Cancer Centre, London, Ontario; 4 Hotel Dieu Hospital, St. Catharines, Ontario; 5 Cross Cancer Institute, Edmonton, Alberta; 6 Toronto Hospital, Toronto, Ontario; 7 Ontario Cancer Treatment and Research Foundation, Toronto–Sunnybrook Regional Cancer Centre, North York, Ontario; 8 Ontario Cancer Treatment and Research Foundation, Kingston Regional Cancer Centre, Kingston, Ontario; 9 Ontario Clinical Oncology Group, Hamilton Ontario; and 10 Knoll Pharma Inc., Markham, Ontario, Canada

Received March 6, 1996

platinum-resistant, further chemotherapy is usually ineffec-

Purpose: Previous phase I and II studies of intraperitoneal recombi- tive and regular paracenteses are often required to control

nant human tumor necrosis factor-a (rhTNF-a) suggested a high de-

gree of efficacy in reducing or eliminating ascitic fluid. To more accu-

rately determine the efficacy of this agent, the role of

versus paracentesis plus intraperitoneal rhTNF-a was studied in a

randomized trial. Patients and methods: Thirty-nine patients

symptomatic ascites with a volume of ú1000 ml from recurrent epithe-

lial ovarian carcinoma or primary peritoneal carcinoma, which was refractory to standard therapy, were randomized either to receive 0.06 mg/m 2 rhTNF-a (Knoll, Canada) (the dose determined optimal from

the symptoms of ascites accumulation. However, in two non- randomized studies in ovarian carcinoma patients with ma-

paracentesis lignant ascites refractory to conventional therapy, recombi-

nant human tumor necrosis factor-a (rhTNF-a) infused intra- with peritoneally resulted in complete disappearance of ascites in

31 of 32 patients, usually with 1 to 3 infusions of rhTNF-a

[1, 2]. To evaluate the efficacy of intraperitoneal (ip) rhTNF- a in such patients, a randomized study comparing paracente-

phase I and II studies) intraperitoneally after drainage of fluid or to sis alone to paracentesis plus ip rhTNF-a was performed. receive drainage alone. A maximum of three treatments were given at weekly intervals. Eighteen patients were randomized to receive PATIENTS AND METHODS

rhTNF-a. Results: None of 18 evaluable rhTNF-a patients had either a complete response (CR) (no clinical evidence of ascites and õ400

ml of fluid on ultrasound) or a partial response (PR)

ascites and £1000 ml of fluid on ultrasound). There were no CRs or

PRs in the 17 evaluable patients who received drainage alone. The intraperitoneal infusion of rhTNF-a was generally well tolerated. Mod- erate to severe toxicity consisted of pain/discomfort in 42.1%, fever/

Thirty-nine patients with histologically proven epithelial (asymptomatic malignancies of the ovary with symptomatic ascites due to

recurrent disease were studied. They had an estimated ascites volume of greater than 1000 ml and had received at least one

course of platinum-based chemotherapy and also fulfilled

chills in 36.9%, nausea/vomiting in 10.5%, edema in 10.5%, and hypo- the following eligibility criteria: (1) life expectancy over 3

tension in 5.3% of patients receiving rhTNF-a. Conclusion: rhTNF- months, (2) age 18 years or older, (3) Karnofsky perfor- a, as given in this study, was not effective in preventing recurrence of mance status §60, (4) no more than two different previous

ascites in this

patient population. 1997 Academic Press chemotherapy regimens, (5) no radiation therapy, major sur- gery, or any form of chemotherapy or immunotherapy within the previous month, (6) serum creatinine £166 mmol/liter,

INTRODUCTION

(7) serum bilirubin, SGOT, and SGPT £2.5 times upper limit of normal, (8) hemoglobin §95 g/liter, leukocytes §3.5 1 10 9 /liter, and platelets §100 1 10 9 /liter, (9) no evidence of other severe nonmalignant disease, (10) no evidence of bowel obstruction, (11) no loculation of ascites or mucinous

Despite optimal surgery and chemotherapy, the majority of patients with advanced epithelial ovarian carcinoma de- velop recurrent disease, often complicated by the develop- ment of malignant ascites. In patients whose tumors are

ascites, and (12) were able to give informed written consent. Glucocorticosteroids and antithrombotic or thrombolytic * Supported by Knoll Pharma Inc. drugs could not be administered during the course of the

0090-8258/97 $25.00 Copyright 1997 by Academic Press All rights of reproduction in any form reserved.

80

TUMOR NECROSIS FACTOR-a IN OVARIAN CARCINOMA

81

study and patients on diuretics or nonsteroidal anti-inflam-

ume were added to give the total ascites volume. Patients

matory medications had to be on a stable dose of these with drainage not performable due to technical reasons had

medications for at least 3 weeks prior to entering the study. A nonrandomized open label protocol was instituted for patients randomized to the control arm who subsequently developed progressive ascites. These patients were subse- quently treated according to the protocol outlined for the patients randomized to the rhTNF-a arm of the randomized protocol. There were a total of 16 patients entered on the

their total fluid estimated by ultrasound. All patients from whom any fluid could be drained received their second treat- ment according to their randomization group. Those patients with no drainable ascites proceeded to follow-up and were reexamined 3 weeks after their assessment (Day 29) (see Fig. 1). This procedure was repeated 1 week following the second treatment (Day 15). All patients from whom any fluid

open label protocol. could be drained received their third treatment according to their treatment allocation. Patients with no drainable ascites

Drug Dosage and Administration

Patients with clinically symptomatic ascites fulfilling the entry criteria were randomly assigned to one of two different treatment modalities (see Fig. 1): Group A, paracentesis fol- lowed by ip administration of rhTNF-a; and Group B, para- centesis alone. Patients’ weight and abdominal circumfer- ence were noted and a catheter was placed into the abdominal

proceeded to follow-up and were reexamined 3 weeks after their assessment (Day 36) (see Fig. 1). Patients having re- ceived a third treatment were reexamined 4 weeks after this treatment (Day 43) (see Fig. 1). All patients were reexamined at the end of their follow-up period (4 weeks after their last treatment). Again, patients were examined for evidence of ascites, both clinically and by ultrasound. Patients with fluid present on ultrasound were drained and the residual fluid

cavity (by ultrasonic guidance if required). Ascites was volume was estimated by ultrasound. Patients in whom

drained as completely as possible (either by gravity or using a vacuum bottle), and the volume drained was recorded. The

drainage was not performable for technical reasons had their total fluid volume estimated by ultrasound. Patients were

volume of residual ascites was then estimated by ultrasonog- then assigned to the response or failure group according

raphy using the recommendations of Goldberg [3] (see ultra- sound response criteria below). Patients assigned to Group A then received an infusion of rhTNF-a via the catheter

to the definitions given below. A clinical assessment was conducted by the investigator and a second physician not involved in the trial. Patients not responding to therapy after

used for drainage. Lyophilized rhTNF-a was provided by three treatments or with a relapse within their follow-up

Knoll Pharma Inc. (Markham, Ontario). The contents of a rhTNF-a vial were diluted to 1.0 ml in sterile water for injection and the patient’s specific dose (0.06 mg rhTNF- a/m 2 BSA) was further diluted in 1000 ml normal saline containing 0.5% (i.e., 25 ml of 20%) human serum albumin to prevent nonspecific adhesion of rhTNF-a to the infusion system. The solution was infused over 2 hr. In order to ensure a uniform dispersion of rhTNF-a solution, the infusion took place with patients in the supine position and patients were

turned from side to side every 30 min for 2

hr postinfusion.

Group B patients received only paracentesis. Patients were hospitalized on the day of treatment and remained under clinical surveillance for at least 6 hr after the start of therapy. As prophylaxis against the known flu- like side effects of rhTNF-a, oral acetaminophen (650 mg) was administered 20 min prior to treatment. Nausea and/or vomiting were controlled by the therapeutic use of com- monly used antiemetics (excluding glucocorticosteroids). Hypotension was treated with normal saline infusions or inotropic drugs, when deemed necessary. Patients were treated at weekly intervals up to a maximum of three treatments (see Fig. 1). One week after the first treatment (Day 8), patients were reexamined. Patients were examined for evidence of ascites both clinically and by ultra-

period could receive other therapy at the investigator’s dis- cretion once evidence of treatment failure had been docu- mented. Otherwise, no additional chemotherapy was allowed until the end of the follow-up period. Patients were clinically reevaluated for signs of delayed response or relapse of asci- tes at 3 and 6 months after their last cycle of therapy.

Evaluation of Efcacy

To be evaluable for efficacy, patients were examined 4 weeks after the last treatment or at the time of relapse of the ascites. Patients who were not evaluable at 4 weeks after last treatment (i.e., death or loss to follow-up) and with no information about a relapse were regarded as lost cases and evaluated separately. The time between the last treatment to 4 weeks after the last treatment was defined as the ‘‘follow- up period.’’

Response Criteria (i) Clinical response.

(1) Complete response—Clinical (CRc): Lack of accu- mulation of ascites at the end of the follow-up period as determined by clinical observation.

sound. Any patient with fluid present on ultrasound had the (2) Partial response—Clinical (PRc): Minimal fluid re-

fluid drained. Residual fluid was then estimated by ultra- sound. The drained volume plus the estimated residual vol-

currence, at the end of the follow-up period as determined by clinical observation (asymptomatic ascites).

82

HIRTE ET AL.

FIG. 1 Treatment schema.
FIG. 1
Treatment schema.

In order to minimize bias, a clinical assessment (abdomi- nal girth measurement, assessment for the presence of ascites and estimation of ascites fluid volume) was conducted at baseline and at the end of the follow-up period (or at time of relapse) by a second physician who was unaware of the treatment received by the patient. In the event of a discrep- ancy between the trial investigator and the second physician,

detect subdiaphragmatic fluid, (b) longitudinal scans of the liver through the left lobe and through the mid and lateral portions of the right lobe, (c) inspection of the abdominal cavity by three transverse scans across the abdomen (approx- imately 5 cm above, across, and approximately 5 cm below the diaphragm), and (d) scanning of the space of Douglas (standard technique or intravaginal ultrasound).

the assessment of the second physician was used to deter- The sonographic criteria used to determine ascites volume

mine response.

(ii) Ultrasound response.

Ultrasound examinations were

performed in accordance with the following guidelines: (a) A transverse scan through the cranial sections of the liver to

were as described by Goldberg [3]: (a) fluid accumulation was termed as none if there were no signs of fluid in the abdominal cavity; (b) fluid accumulation was termed as mini- mal ascites (1–400 ml) if there were occasional thin layers

TUMOR NECROSIS FACTOR-a IN OVARIAN CARCINOMA

83

of fluid between single bowel loops or small fluid collections in the space of Douglas; (c) fluid accumulation was termed as minor ascites (401–1000 ml) if there were thin layers of fluid around the liver (below the diaphragm and the interior edge) or frequently occurring layers of fluid between bowel loops; and (d) fluid accumulation was termed as major

TABLE 1 Baseline Patient Characteristics

Factor

TNF

Paracentesis

P value

Mean age { SD Mean No. of days diagnosis

59.9 { 9.1

61.5 { 9.1

0.59

(ú1000 ml) if there was massive fluid collection around the to randomiziation { SD 817.3 { 1238 798.9 { 582 0.60

liver or gross collections of abdominal fluid with floating bowel loops. Ultrasound responses were defined as follows. (1) Com-

Mean Karnofsky { SD Mean chemotherapy courses { SD Mean previous paracenteses

Stage of disease (%)

II

III

IV

Unknown

71.9 { 15.3

75.7 { 9.8

0.36

2.1 { 0.8

2.2 { 0.7

0.73

3.8 { 2.6

4.0 { 2.8

0.85

 

0.20

0 (0.0)

2 (9.5)

17 (94.4)

15 (71.4)

1 (0.0)

3 (14.3)

0 (0.0)

1 (4.8)

plete response—Ultrasound (CRu): No or minimal accumu- { SD

lation of ascites (õ400 ml) at the end of the follow-up period as determined by drainage whenever medically feasible, and residual volume estimated by ultrasound. When drainage

was not medically feasible, total volume was determined

by

ultrasound alone. (2) Partial response—Ultrasound (PRu):

Minor fluid recurrence (£1000 ml) at the end of the follow-

up period as determined by drainage whenever medically

Toxicity Evaluation

feasible, and residual volume estimated by ultrasound. When The WHO criteria for toxicity were used for the grading

drainage was not medically feasible, the total volume was determined by ultrasound alone. The ultrasound evaluation was conducted by a designated ultrasonographer at each center. Whenever possible, this in- dividual was unaware of the treatment received by the pa- tient.

Definition of Response

A responder was defined as a patient who met one of the following criteria: (a) A complete response required both a

clinical complete response as well as an ultrasound

complete

response. (b) A partial response required (i) a complete clini- cal response and a partial ultrasound response (CRc / PRu)

or (ii) a partial clinical response and a complete ultrasound

response (PRc / CRu) or (iii) a partial clinical response and

of severity of symptoms and abnormal ndings which were

as follows: Grade 1 Å mild toxicity, Grade 2 Å moderate

toxicity, Grade 3 Å severe toxicity, and Grade threatening toxicity.

4

Å

life-

Quality of Life Assessment

To assess the effect of treatment on quality of life, the EORTC Core Quality of Life Questionnaire was adminis- tered at baseline before the first treatment, immediately prior to each additional treatment, and 4 weeks after the last treat- ment or within 4 weeks of relapse.

Statistical Analysis

Student’s t test was used to determine the significance of differences for all comparisons, except for stage of disease, where a x 2 test was used.

a partial ultrasound response (PRc / PRu). Trial Execution

The trial commenced in June 1992. Although it was deter-

Denition of Relapse mined a priori that a sample size of 128 patients (64 patients

Relapse was defined as the occurrence of symptomatic ascites requiring aspiration or fluid volume greater than 1000 ml as determined by drainage and/or ultrasound, which oc- curred after the last scheduled treatment but prior to the end of the 4-week follow-up period.

Definition of Treatment Failure

Any of the following criteria constituted treatment failure:

per arm) was required to detect at least a 30% difference in response rates between treatment arms (this was believed to be a clinically relevant difference), in the fall of 1993, after 39 patients had been entered on the study, investigators be- came concerned by the absence of any responses in the 18 patients treated with rhTNF-a. This observation was verified by the trial coordination center, and at a subsequent meeting of investigators in November 1993, it was decided to close the trial to further accrual.

(a) Patients who relapsed prior to the end of the 4-week

RESULTS

follow-up period, (b) patients who, at the end of the 4-week follow-up period, did not meet the definition of a responder, A total of 39 patients were randomized: 18 to rhTNF-a

or (c) patients who, at anytime, did not respond to therapy and were removed from the study because of adverse reac- tions.

and 21 patients to paracentesis alone. There were no signifi- cant differences in baseline characteristics of patients be- tween treatment groups (Table 1). All patients completed at

84

HIRTE ET AL.

least one treatment (paracentesis { injection of rhTNF-a), DISCUSSION

15

patients on the treatment arm and 19 patients on the

control arm completed at least two treatments, and 14 pa- Phase I and II studies of intraperitoneal rhTNF-a sug-

tients on the treatment arm and 17 on the control arm com- pleted the maximum of three treatments. Four patients failed to complete the study and were deemed inevaluable. All were on the control arm. One patient had two treatments and then expired with no ultrasound or drainage performed prior to this. Two patients developed bowel obstruction; one required a colostomy after three treatments but no mention was made of whether ascites was present at that time and a second patient developed bowel obstruction after one treat- ment but there was no indication of whether ascites was present because no ultrasound or drainage was performed. A fourth patient had three treatments but was too ill to be examined for entry into the open label protocol and no infor-

gested substantial efficacy in controlling malignant ascites due to recurrent ovarian carcinoma [1, 2]. The phase I trial

indicated that 0.08 mg/m

  • 2 rhTNF-a was both safe and effec-

tive in the treatment of malignant ascites [1]. This dose was subsequently chosen for a phase II multicenter trial [2]. Twenty of 20 ovarian carcinoma patients treated had a response. The present study was performed to confirm this activity in a randomized setting in a similar population of ovarian carcinoma patients using the same dose of rhTNF-

a. However, in our study, intraperitoneal rhTNF-a showed no evidence of efficacy in the patients studied. In contrast, in the phase I and II studies using the same source of rhTNF- a, responses were seen in 30 of 31 patients with ascites due

mation about ascites was obtained. to recurrent ovarian carcinoma [1, 2].

One rhTNF-a patient had a complete clinical response; It is not clear why rhTNF-a showed no efficacy in our

however, no ultrasound was performed and thus the patient did not meet the criteria for either a complete or partial response. The patient expired 10 days later. There were no responses in the other 17 patients randomized to the rhTNF- a arm. There were no responses with respect to ascites accu- mulation in 17 evaluable control patients nor were there any responses in the 16 patients entered on the open label study after developing recurrent ascites in the control arm. No reduction in measurable solid tumor masses were noted. Be- cause there were no documented responses in patients treated with rhTNF-a in the randomized study or the open label study, it was decided to close the study to further accrual at this point. There was a significant difference in adverse events be- tween treatment arms with 16 of 18 (88.9%) of patients on the treatment arm developing at least one adverse event, but only 3 of 21 (14.3%) patients on the control arm (P õ 0.0001). There were a total of 62 adverse events on the treatment arm (1.3 adverse events per treatment) and 4 in the control arm (0.070 adverse events per treatment). The most common adverse events in the treatment arm were fever/fatigue/chills (0.45 events per treatment), nausea/vom- iting (0.36 events per treatment), and pain/discomfort at the paracentesis/injection site (0.41 events per treatment). Ad- verse events in the control arm were limited to fever/fatigue/ chills (0.036 events per treatment) and pain/discomfort at the paracentesis site (0.036 events per treatment). Most adverse

study. In the phase I and II studies, patient selection and response criteria were similar and the dose of rhTNF-a

ranged from 20 to 350 mg/m 2 [1, 2]. The dose of rhTNF-a

used in our study (0.06 mg/m

  • 2 ) had equivalent activity

to

0.08 mg/m 2 used in the phase II study. Although the lot of

rhTNF-a used in our study was different from that in the phase I and II studies, the rhTNF-a used in our study showed appropriate TNF-a content and biological activity as mea- sured by the L-929 cytotoxicity assay (4). Thus the bioactiv- ity of the rhTNF-a used in the study did not appear to have influenced the result of the study. In a nude mouse model of intraperitoneally xenografted human ovarian carcinoma, rhTNF-a in doses of 1 to 5 1 10 4 units/mouse/day prevented the development of ascitic disease when intraperitoneal ther- apy started the day after tumor cell injection. If treatment was delayed for 7 days, only the highest doses of rhTNF-a showed any activity (5). There are no other animal or clinical data available indicating whether there is a dose–response relationship for the administration of intraperitoneal TNF-a and ascites resolution. Because of the high degree of efficacy

of the 0.08 mg/m 2 dose in the phase II study, there was no provision to escalate the dose of rhTNF-a in the initial trial design, and given the complete lack of activity of the 0.06

mg/m

  • 2 dose in the first 39 patients treated in our study (which

had been effective in all 20 ovarian carcinoma patients treated in the phase II study), the dose of rhTNF-a was not

escalated.

events (95%) were mild or moderate in degree of severity In the fall of 1993, a decision was made to terminate

(Table 2). There were no significant differences in survival. Patients on the treatment arm survived a mean of 132.4 { 80.3 days after randomization, while patients on the control arm sur-

patient accrual to this study based on an interim analysis of outcome. This analysis was not a planned interim analysis but was performed in response to concerns raised by study investigators. The decision to terminate accrual to any study

vived 145.8 { 161.9 days (P Å 0.58). is a difficult

one. However, the study steering committee

Because there were no responses in patients treated with rhTNF-a, the quality of life data collected on these patients

believed that because no efficacy had been demonstrated in the first 18 patients randomized to the rhTNF-a arm of the

were not further analyzed. study, it would be unethical to continue patient accrual.

TUMOR NECROSIS FACTOR-a IN OVARIAN CARCINOMA

TABLE 2 Severity of Adverse Events

85

 

TNF patients

 

Paracentesis patients

 
 

Adverse

Adverse

event

event

Adverse event

Severity

n

treatment

n

treatment

P value

Fever/fatigue/chills

Mild

11

0.234

1

0.018

Moderate

9

0.192

1

0.018

õ0.001

Severe

1

0.021

0

0.000

Nausea/vomiting

Mild

13

0.277

0

0.000

Moderate

4

0.085

0

0.000

õ0.0001

Severe

0

0.000

0

0.000

Pain/discomfort

Mild

3

0.064

0

0.000

Moderate

14

0.298

1

0.018

õ0.0001

Severe

2

0.043

1

0.018

Hypotension

Mild

0

0.000

0

0.000

Moderate

1

0.021

0

0.000

0.45

Severe

0

0.000

0

0.000

Edema

Mild

0

0.000

0

0.000

Moderate

2

0.043

0

0.000

0.20

Severe

0

0.000

0

0.000

Other GI

Mild

0

0.000

0

0.000

Moderate

0

0.000

0

0.000

1.0

Severe

0

0.000

0

0.000

Other

Mild

1

0.021

0

0.000

Moderate

1

0.021

0

0.000

0.20

Severe

0

0.000

0

0.000

Total: All events

62

1.319

4

0.070

The management of ascites due to recurrent ovarian carci-

crine fashion by eliciting fluid leakage from the neighboring

noma is problematic. Paracentesis alone is the technique host vasculature [24, 25]. Factors which have been demon-

most commonly used to palliate symptoms of ascites. Perito- neal shunts have been used to palliate refractory ascites in ovarian carcinoma patients with limited success and have been associated with high complication rates including shunt blockage and thrombotic and embolic complications related

strated to mediate such effects include vascular permeability factor (VPF) [26], epidermal growth factor (EGF), and transforming growth factor-a (TGF-a) [27]. In addition, the

neovasculature of a growing tumor is thought to be intrinsi- cally leaky, thereby providing an additional site for fluid

to the shunt [6–9]. extravasation from the vascular space [28].

A variety of other agents have been administered intraperi- Significantly elevated concentrations of VPF have been

toneally, in an attempt to control malignant ascites. These include radioisotopes (as well as radiolabeled monoclonal antibody) [10–12], chemotherapeutic agents (bleomycin) [13–15], and biological response modifiers (Cornebacterium

demonstrated in malignant ascites of patients with ovarian carcinoma [29] and the malignant tissue appears to be one of the sources of this activity [30]. In an animal model of ascites-producing tumors, VPF accumulated in the vessels

parvum, interferon-b, interferon-a, and interleukin-2 with or of the tumor and the peritoneal surfaces [31, 32]. VPF may

without LAK cells) [16–22]. Although reduction in ascites accumulation was achieved in 33 to 87% of patients, re- sponses were often only of short duration. In addition, these studies were either phase I or II in design and the true test

also be a critical factor in supporting the widespread intra- peritoneal growth of ovarian carcinoma by allowing the in- duction of neovascularization and stroma formation of meta- static deposits [33]. Indeed, injection of anti-VPF antibodies

of any therapy is in a randomized setting. in nude mice transplanted with human tumors led to tumor

The development of ascites in patients with ovarian carci- noma contributes significantly to the morbidity of this dis-

regression and decreased density of the tumor-supporting vasculature [34].

ease. Mean survival after the discovery of ascites is 30–40 EGF and TGF-a have also been shown to induce the

weeks. Although early work emphasized the importance of lymphatic obstruction by tumor cells (23), recently new in- sight has been shed on the pathophysiology of malignant ascites. Secreted products of tumor cells may act in a para-

production of ascites when injected intraperitoneally in a murine model [27]. The mechanism by which EGF or TGF- a stimulates ascites formation is not clear. Indomethacin, a blocker of prostaglandin (PG) synthesis, inhibited the effect

86

HIRTE ET AL.

of EGF, suggesting that PGs, especially PGE 2 and PGI 2 , which have been reported to increase vascular permeability, are involved in EGF-stimulated ascites formation [27]. How- ever, the effect of indomethacin may also have been medi- ated by some compensatory or other unknown mechanism. Since both EGF and TGF-a are angiogenic mediators and induce discontinuity of the basement membrane, they may also induce the production of bloody ascites through the formation of fragile new blood vessels. The ascites from patients in the present study was not assessed for levels of EGF, TGF-a, or VPF. Although it is possible that EGF,

TGF-a, or VPF may have inhibited the effects of the admin- istered TNF-a, similar effects (i.e., lack of activity) should have been noted in the patients in the phase I and II studies. However, the rhTNF-a was highly effective in reducing asci- tes volumes in these earlier studies, suggesting that these other factors, which would have been expected to be present,

were not able to abrogate the effects of rhTNF-a.

Although

soluble TNF receptors (s-TNF-R) have been shown to be present in malignant ascites [35], their role in affecting the response to rhTNF-a administered intraperitoneally is not clear. Levels of s-TNF-R in serum or ascites of patients in the present study were not determined. s-TNF-R would have

been expected to be present in the ascites of patients treated in the phase I and II studies as well as in the ascites of patients in the present study. The fact that patients in the phase I and II studies had very high response rates, whereas there were no responses in the present study, suggested that s-TNF-R were not likely responsible for this variability in response to rhTNF-a. It would appear that ultimately approaches to reverse un- derlying pathophysiology of ascites generation, either by blocking production or interfering with the effects of factors

such as VPF or by more effectively treating the

underlying

tumor, are most likely to be effective in controlling accumu-

lation of malignant ascites.

REFERENCES

  • 1. Raeth, U., Kaufmann, M., Schmid, H., Hofmann, J., Wiedenmann, B., Kist, A., Kempeni, J., Schlick, E., Bastert, G., Kommerell, B., and Maennel, D. Effect of intraperitoneal recombinant human tumor necro- sis factor alpha on malignant ascites, Eur. J. Cancer 27, 121–125 (1991).

  • 2. Raeth, U., Schmid, H., Karck, U., Kempeni, J., Schlick, E., and Kauf- mann, M. Phase II trial of recombinant human tumor necrosis factor- a (rHuTNF) in patients with malignant ascites from ovarian

carcinomas

and non-ovarian tumors with intraperitoneal spread, Proc. Am. Soc.

Clin. Oncol. 10, 187 (1991).

  • 3. Goldberg, B. Evaluation of ascites by ultrasound, Radiology 96, 15– 22 (1970).

  • 4. Knoll Technical Report 000981-02-D, on file, Knoll AG, Ludwigsha- fen, Germany.

  • 5. Balkwill, F. R., Ward, B. G., Moodie, E., and Fiers, W. Therapeutic potential of tumor necrosis factor-a and g-interferon in experimental

  • 6. Cheung, D. K., and Raaf, J. H. Selection of patients with malignant ascites for a peritoneovenous shunt, Cancer 50, 1204–1209 (1982).

  • 7. Lacy, J. H., Wieman, T. J., and Shively, E. H. Management of malig- nant ascites, Surg. Gynecol. Obstet. 159, 397–412 (1984).

  • 8. Souter, R. G., Wells, C., Tarin, D., and Kettlewell, M. D. Surgical and pathological complications associated with peritoneovenous shunts in management of malignant ascites, Cancer 55, 1973–1975 (1985).

  • 9. Edney, J. A., Hill, A., and Armstrong, D. Peritoneovenous shunts palli- ate malignant ascites, Am. J. Surg. 158, 598–601 (1989).

    • 10. Ariel, I. M., Oropeza, R., and Pack, G. T. Intracavitary administration of radioactive isotopes in the control of effusions due to cancer, Cancer 19, 1096–1102 (1966).

11. Jackson, G. L., and Blosser, N. M. Intracavitary chromic phosphate ( 32 P) colloidal suspension therapy, Cancer 48, 2596–2598 (1981).

12. Buckman, R., De Angelis, C., Shaw, P., Covens, A., Osborne, R., Kerr, I., Reed, R., Michaels, H., Woo, M., Reilly, R., Law, J., Baumal, R., Groves, E., and Marks, A. Intraperitoneal therapy of malignant ascites associated with carcinoma of ovary and breast using radioiodinated monoclonal antibody 2G3, Gynecol. Oncol. 47, 102–109 (1992).

13. Bitran, J. D. Intraperitoneal bleomycin: Pharmacokinetics and results of a phase II trial, Cancer 56, 2420–2423 (1985).

14. Ostrowski, M. J. An assessment of the long-term results of controlling the reaccumulation of malignant effusions using intracavitary bleomy- cin, Cancer 57, 721–727 (1986).

15. Paladine, W., Cunningham, T. J., Sponzo, R., Donavan, M., Olson, K., and Horton, J. Intracavitary bleomycin in the management of malignant effusions, Cancer 38, 1903–1908 (1976).

  • 16. Mantovani, A., Sessa, C., Peri, G., Allavena, P., Introna, M., Polentaru- tti, N., and Mangioni, C. Intraperitoneal administration of corynebacte- rium parvum in patients with ascitic ovarian tumors resistant to chemo- therapy: Effects on cytotoxicity of tumor-associated macrophages and NK cells, Int. J. Cancer 27, 437446 (1981).

17. Rambaldi, A., Introna, M., Colotta, F., Landolfo, S., Colombo, N., Mangioni, C., and Mantovani, A. Intraperitoneal administration of inter- feron b in ovarian cancer patients, Cancer 56, 294–301 (1985).

  • 18. Gebbia, V., Russo, A., Gebbia, N., Valenza, R., Testa, A., Palmeri, S., and Rausa, L. Intracavitary beta-interferon for the management of pleu- ral and/or abdominal effusions in patients with advanced cancer refrac- tory to chemotherapy, In Vivo 5, 579–582 (1991).

19. Bezwoda, W. R., Seymour, L., and Dansey, R. Intraperitoneal recombi- nant interferon-alpha 2b for recurrent malignant ascites due to ovarian cancer, Cancer 64, 1029–1033 (1989).

  • 20. Stuart, G. C. E., Nation, J. C., Snider, D. D., and Thunberg, P. Intraperi- toneal interferon in the management of malignant ascites, Cancer 71, 20272030 (1993).

  • 21. Lissoni, P., Barni, S., Ardizzoia, A., Paolorossi, F., Tisi, E., Crispino, S., and Tancini, G. Intracavitary administration of interleukin-2 as pal- liative therapy for neoplastic effusions, Tumori 78, 118–120 (1992).

  • 22. Li, D-J., Wang, Y-R., Tan, X-Y., Wang, H-Z., Yao, X-D., and Ba, D-N. A new approach to the treatment of malignant effusion, Chin. Med. J. 103, 998–1002 (1990).

  • 23. Holm-Nielson, P. Pathogenesis of ascites in peritoneal carcinomatosis, Acta Pathol. Microbiol. Scand. 33, 10–21 (1953).

  • 24. Hirabayashi, K., and Graham, J. Genesis of ascites in ovarian cancer,

Am. J. Obstet. Gynecol. 106, 492–497 (1970). 25. Garrison, R. N., Heuser, L. S., and Galloway, R. H. Mechanisms of malignant ascites production, J. Surg. Res. 42, 126–132 (1987). 26. Senger, D. R., Galli, S. J., Dvorak, A. M., Peruzzi, C. A., Harvey, V. S.,

human ovarian cancer, Cancer Res. 47, 4755–4758 (1987). and Dvorak, H. F. Tumor cells secrete a vascular permeability factor

TUMOR NECROSIS FACTOR-a IN OVARIAN CARCINOMA 87

that promotes accumulation of ascites fluid, Science 219, 983–985

(1982).

  • 27. Ohmura, E., Tsushima, T., Kamiya, Y., Okada, M., Onoda, N., Shi- zuma, K., and Demura, H. Epidermal growth factor and transforming growth factor a induce ascitic fluid in mice, Cancer Res. 50, 4915– 4917 (1990).

  • 28. Heuser, L. S., Taylor, S. H., and Folkman, J. Prevention of carcinoma- tosis and bloody malignant ascites in the rat by an inhibitor of angiogen- esis, J. Surg. Res. 36, 244–250 (1984).

  • 29. Yeo, K-T., Wang, H. H., Nagy, J. A., Sioussat, T. M., Ledbetter, S.

R.,

Hoogewert, A. J., Zhou, Y., Masse, E. M., Senger, D. R., and Dvorak,

H. F. Vascular permeability factor (vascular endothelial growth factor) in guinea pig and human tumor and inflammatory effusions, Cancer Res. 53, 2912–2918 (1993).

Sotrel, A., Manseau, E. J., Van De Water, L., and Senger, D. R. Distri- bution of vascular permeability factor (vascular endothelial growth fac- tor) in tumors: Concentration in tumor blood vessels, J. Exp. Med. 174, 1275–1278 (1991).

  • 32. Nagy, J. A. Immunochemical localization of vascular permeability fac- tor (vascular endothelial growth factor) in ascites tumors: distribution in peritoneal wall microvasculature, J. Cell Biol. 115, 264a (1991).

  • 33. Nagy, J. A., Brown, L. F., Senger, D. R., Lanir, N., Van De Water, L., and Dvorak, H. F. Pathogenesis of tumor stroma generation: A critical role for leaky blood vessels and fibrin deposition, Biochim. Biophys. Acta 948, 305–326 (1988).

  • 34. Kim, K. J., Li, B., Winer, J., Armanini, M., Gillett, N., Phillips, H. S., and Ferrara, N. Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumor growth in vivo, Nature 362, 841–844 (1993).

  • 30. Olson, T. A., Mohanraj, D., Carson, L. F., and Ramakrishnan, S. Vascu- lar permeability factor gene expression in normal and neoplastic human

  • 35. Andus, T., Gross, V., Holstege, A., Ott, M., Weber, M., David, M., Gallati, H., Gerok, W., and Schoelmerich, J. High concentrations of

ovaries, Cancer Res. 54, 276280 (1994). soluble tumor necrosis factor receptors in ascites, Hepatology 16, 749–

  • 31. Dvorak, H. F., Siousset, T. M., Brown, L. F., Berse, B., Nagy, J. A., 755 (1992).