Annals of Oncology 18: 945–949, 2007

original article doi:10.1093/annonc/mdl499

Published online 13 February 2007

Pattern and prognostic factors in patients with

malignant ascites: a retrospective study
A. A. Ayantunde* & S. L. Parsons
Department of Surgery, Nottingham University Hospitals, Nottingham City Hospital Campus, Nottingham, UK

Received 10 August 2006; revised 3 November 2006 and 11 December 2006; accepted 15 December 2006

Background: Malignant ascites is a manifestation of end stage events in a variety of cancers and associated with
a poor prognosis. We evaluated the pattern of cancers causing malignant ascites and factors affecting survival.
Patients and methods: Patients coded with the International Classification of Diseases-9 coding system
for malignant ascites over a 2-year period were reviewed. The clinicopathological data and patients’ survival

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 16, 2014

were compared among cancer groups.
Results: There were 209 patients (140 females and 69 males), median age being 67 (30–98) years. The commonest
cancer was ovarian followed by gastrointestinal (GI) cancers. Fifty-eight per cent of the patients had symptoms related
to the ascites. Liver metastases were significantly commoner in the GI cancers (P = 0.0001). Fifty-four per cent of
our patients presented with ascites at the initial diagnosis of their cancer. Paracentesis was given to 112, diuretics
to 70 and chemotherapy to 103 patients. The median survival following diagnosis of ascites was 5.7 months. Ovarian
cancer favoured longer survival while low serum albumin, low serum protein and liver metastases adversely affected
survival. The independent prognostic factors for survival were cancer type, liver metastases and serum albumin.

original
Conclusion: The identified independent prognostic factors should be used to select patients for multimodality

article
therapy for adequate palliation.
Key words: cancers, chemotherapy, malignant ascites, paracentesis, prognotic factors, survival

introduction with increasing availability and use of appropriate and potent

combination chemotherapy.
Malignant ascites is a manifestation of end stage events in The objectives of this study were to determine the pattern
a variety of cancers and is associated with significant morbidity. of cancers causing malignant ascites, treatment modalities
It accounts for about 10% of all cases of ascites and usually offered and factors affecting survival.
caused by ovarian, endometrial, breast, oesophageal, gastric,
colorectal, lung, pancreatic, hepatobiliary and primary
peritoneal carcinomas [1–3]. It is known that about 50% of
patients with malignant ascites present with ascites at the initial patients and methods
diagnosis of their cancer [4, 5]. The onset and progression of The medical records of all patients coded for malignant ascites using the
malignant ascites is associated with deterioration in quality of International Classification of Diseases-9 coding system from January 2003
life (QoL) and a poor prognosis. There are, however, no to December 2004 at the Nottingham City Hospital were reviewed. The
generally accepted evidence-based guidelines for evaluation and patients included in this study had clear documentation of ascites of
treatment of this condition. There are also no clinical predictors malignant origin in their clinical notes. The development of ascites in
that identify cancer patients who will develop this distressing patients with history of either intra-abdominal or extra-abdominal
entity; hence there are no preventive measures for its malignancy in the absence of liver cirrhosis and other causes of
development. nonmalignant ascites was accepted as being malignant. Malignant origin
of the ascites in these patients was usually confirmed by one or more of
There are different approaches to the treatment of malignant
cytological examination, imaging, laparoscopy or laparotomy.
ascites from symptomatic relief with simple drainage procedures
The data collected included demographics, cancer entity, time interval
to chemotherapy and debulking surgery aimed at treating the
between the diagnosis of the primary cancer and that of the ascites,
underlying cancer. It is envisaged that QoL and possibly the symptoms of ascites, method of diagnosis of ascites, laboratory parameters
survival of patients with malignant ascites may be improved at diagnosis, the presence of metastases at diagnosis of ascites, treatment of
ascites and survival.
*Correspondence to: Dr A. A. Ayantunde, Department of Surgery,
Nottingham University Hospitals, Nottingham City Hospital Campus, Nottingham,
Statistical analysis was by the SPSS version 13.0 (SPSS, Chicago, IL,
NG5 1PB, UK. Tel: +44-115-969-1169; Fax: +44-115-962-7764; USA) using Mann–Whitney U-test and log-rank test with P value <0.05
E-mail: [email protected] considered significant.

ª 2007 European Society for Medical Oncology

original article Annals of Oncology

results The median total serum protein and albumin at diagnosis of the
ascites in the 209 patients were 61 (30–78) g/l and 28 (15–43)
There were 209 patients with malignant ascites during this g/l, respectively. There is a significant correlation between the
period with 140 (67%) females and 69 (33%) males. The median presence of liver metastases and low total serum protein
age at diagnosis of the ascites was 67 (30–98) years. The cancer (P = 0.018) but no relationship was found with low serum
types causing the ascites are shown in Figure 1. The proportion albumin (P = 0.390).
of patients developing ascites from each main cancer group The most common treatment offered for the ascites was serial
expressed as a percentage of the total number of patients paracentesis in 112 (53.6%) patients with mean of two (1–7)
registered at the tumour specific multidisciplinary team drainages required. Sixty-three out of 209 (30%) patients had
meetings include 18.3% for gastric cancer, 4.0% for oesophageal the initial drainage of their ascites either at laparoscopy or
cancer, 3.7% for colorectal cancer, 36.7% for ovarian cancer, laparotomy with only 15 out of 63 patients requiring further
3.0% for breast cancer and 21% for pancreaticobiliary cancers. paracentesis. Forty-nine out of 209 (23.4%) patients had no
One hundred and twenty-two (58%) of the patients presented need for a drainage procedure. These patients who had drainage
with symptoms related to the ascites including abdominal at surgery or required no drainage at all during their hospital
swelling (55%), abdominal pain (53%), nausea (37%), anorexia episode had subclinical ascites that were either diagnosed
(36%), vomiting (25%), fatigue (17%), dyspnoea (11%), early radiologically or incidentally found at laparoscopy or
fullness (6%), weight change (5%), ankle swelling (3%) and laparotomy. None of the patients in this series had either
heartburn (1%). The median time interval between the continuous catheter drainage or peritoneovenous shunt
diagnosis of the cancer of origin and that of the ascites was 0.87 placement. Diuretics were used in 70 patients with

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 16, 2014

(0–341) months. In fact, 54% of patients had their ascites at first spironolactone, an aldosterone antagonist used in 56 (80%),
diagnosis of their cancer and the median of 0.87 (not zero) frusemide in six (8.6%) and combination of both in eight
months represents a delay in diagnostic investigations in a few (10.4%). Diuretics tended to be used in patients with liver
patients. Forty-one per cent of the patients had their ascites metastasis (55/56). Intravenous or oral single or combination
diagnosed clinically and those with subclinical ascites, diagnosis chemotherapy was given in 103 (49.3%) patients (41 ovarian,
was made at laparoscopy, laparotomy or with imaging 35 GI, 11 breast, 12 ‘other’ cancers and four unknown primary).
investigations [computed tomography (CT) and ultrasound The commonest chemotherapy in ovarian cancer was a
scans]. carboplatin-based regimen. Intraperitoneal (i.p.) cisplatin was
The sites of metastases at diagnosis of the ascites as used in one patient. The remaining 50.7% of the patients
demonstrated by diagnostic imaging techniques included were entered into the palliative care pathway and had only
peritoneal (181 patients), liver (56 patients), bone (26 patients) symptomatic control of their malignant disease process.
and lung (16 patients). Liver metastases was commonest in Follow-up was until death or at least 17 months after
patients with cancers of gastrointestinal (GI) origin at 45%, diagnosis of ascites. There were 23 patients who were still alive
followed by breast cancer 32%, unknown cancer 11%, ovarian at the end of the study (follow-up range 17–52 months). The
cancer 7% and ‘other’ cancer 5%. The difference in the overall median survival after the diagnosis of malignant ascites
incidence of liver metastases in the tumour groups is statistically was 5.7 (95% confidence interval 3.54–7.93) months. Eighty-
significant (P = 0.0001). Bone metastases were most common in eight out of 209 patients had subclinical ascites and these had
patients with a breast primary. Immunocytochemistry of the a slightly longer median survival than patients with clinical
ascites was positive in 140 (67%), suspicious in 10 (4.8%), ascites (7 versus 5 months) but this did not reach a statistical
negative for cancer cells in 12 (5.7%) and was not carried out in significant level (P = 0.159). Patients with ovarian cancer had
47 (22.5%). There was no metastasis detected in 10 patients better survival than those with other cancer groups (P = 0.0001).
on imaging techniques. Of these cytology was positive in five, Patients with ascites of GI cancer origin, other cancers and those
suspicious in two, negative in one and not done in two patients. with unknown primary had the worst survival (Figures 2 and 3).
Levels of serum albumin and total protein at diagnosis of ascites
significantly affected survival. Low serum albumin (<30 g/l) and
low serum total protein (<60 g/l) are both associated with poor
Uterine overall survival (P = 0.0001 and P = 0.025, respectively). The
Urinary bladder
Unknown primary presence of liver metastases was associated with worse survival
Primary site of cancer

Soft tissue sarcoma

Prostate in all cancer groups (P = 0.016). Females survived longer than
Primary peritoneal carcinoma
Pancreas
the males (P = 0.0001) but this is due to better prognosis
Ovarian with ovarian and breast cancer groups. There was improved
Oesophageal
Malignant mesothelioma survival in those patients who had chemotherapy (P = 0.0001).
Lymphoma
Lung Multivariate Cox regression analysis showed cancer type, low
Kidney
Gastric
serum albumin levels and the presence of liver metastasis to
Gallbladder be independent prognostic factors for overall survival.
Colorectal
Breast

0 10 20 30 40 50 60 discussion
Frequency
Symptomatic malignant ascites is a significant problem in the
Figure 1. Primary cancer types causing malignant ascites. palliative care setting and associated with a progressively

946 | Ayantunde & Parsons Volume 18 | No. 5 | May 2007

Annals of Oncology original article
institution during the 2-year review, ovarian cancer had the
highest proportion of patients who developed ascites at 37.7%
followed by pancreaticobiliary cancers 21%, gastric cancer
18.3%, oesophageal cancer 4.0%, colorectal cancer 3.7 and
breast cancer 3.0%. We did not encounter any case of
hepatocellular carcinoma causing ascites in this series. This may
be due to low incidence of primary hepatocellular carcinoma
in the UK population. There is a predominance of the female
patient population (140 of 209) similar to previous studies and
this is accounted for by the predominance of the ovarian and
breast cancer entity.
Fifty-four per cent of our patient population presented with
malignant ascites at the initial diagnosis of their cancer and this
is similar to the series previously reported [4, 5]. These patients
were mainly the ovarian and the GI cancer groups, while
Figure 2. Kaplan–Meier survival curve comparing survival in patients
patients with breast cancer tended to develop ascites due to their
with malignant ascites from different cancer groups.
cancers months or years after their primary cancer had been
diagnosed and treated. The number of malignant ascites due to
an unknown primary cancer in the present study was less than

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 16, 2014

2.00
the previous study done by Parsons et al. [5] (8.1% versus
22.6%) from this same centre and other reported incidence of
13–22% [9]. This is probably due to improvements in
determination of tumour of origin by techniques such as
Median survival in years

1.50
imaging, laparoscopy and immunocytochemical analysis. In our
series, 10 patients had no evidence of metastases on imaging.
Seven out of the 10, however, had malignant or suspicious
1.00 cytology and therefore, probably had peritoneal carcinomatosis
not detected on CT. In another two patients, cytology was not
carried out and in one other the cytology was negative. It may
be that the cytologically negative patient also had peritoneal
0.50
carcinomatosis, although Runyon et al. [3] previously showed
that of their total number patients with malignant ascites, 53.3%
had peritoneal carcinomatosis and all of this group had a
0.00 positive cytology indicating a near 100% sensitivity of cytology
Ovarian Breast GIT Other Unknown
cancer cancer cancers cancers primary in patients with peritoneal carcinomatosis.
(52) (33) (81) (26) (17) Twenty-three of our patients are still alive after a minimum
Figure 3. Median survivals in patients with malignant ascites in years by follow-up of 17 months and most of them are in the ovarian
cancer groups. cancer group. Ovarian cancer is usually amenable to debulking
surgery and has a good response to chemotherapy. Our finding
that ascites of ovarian origin has a better median survival than
deteriorating QoL and a poor prognosis. It is, however, believed all other cancer groups agrees with previous studies [4–8].
that with better understanding of the pathophysiology of This also may have been responsible for the better survival
malignant ascites, better diagnostic evaluation and the use of seen in women compared with men. We therefore reinforce the
multimodality therapy, the QoL and survival of these patients conclusion of our previous study [5] that female patients with
may be improved. There are few studies evaluating the natural malignant ascites from cancer of unknown primary should be
history of malignant ascites and the prognostic factors relating aggressively investigated in case they have an underlying ovarian
to survival [4, 6–8] and some authors [6, 7] in different series cancer. Seventy-nine per cent (41 of 52) of our patients with
have demonstrated the predominance of ovarian cancer causing ovarian cancer had chemotherapy which was mainly a
ascites in the patient’s population studied. carbolatin-based regime with a measure of response.
Parsons et al. [5] carried out a 2-year retrospective review We have evaluated various factors that affect survival in
from this centre over a decade ago in which they showed that patients with malignant ascites and our findings are in keeping
ovarian cancer was the commonest cause of ascites with far with the few published articles addressing this issue [5–8].
better prognosis than patients with GI cancers. We wanted to see Patients with GI cancer as well as those with an unknown
whether there had been any changes in relation to pattern of primary have a very poor survival compared with the ovarian or
cancer causing ascites, treatment options, prognostic factors and breast cancer groups. Overall, we noticed that patients in our
survival in the last 10 years. Again, the commonest cancer of series had longer median survival periods in all cancer groups
origin leading to malignant ascites in our series was ovarian than reported 10 years ago. This may be explained by increasing
representing 25% of the total patient population. Of the total use of aggressive multimodality therapy including combination
cancer cases seen by each multidisciplinary team in our chemotherapy (49% received chemotherapy in the current

Volume 18 | No. 5 | May 2007 doi:10.1093/annonc/mdl499 | 947

original article Annals of Oncology

study compared with 41% in our previous study). Furthermore, Therefore, further studies are required in order to identify those
the success of chemotherapy has improved over the last 10 years patients who will benefit from this therapy. Only one patient
in breast and ovarian cancers. The presence of liver metastases at in this study had i.p. cisplatin therapy and none of the
the time of ascites diagnosis was a significant predictor of poor patients reviewed had intracavitary radiotherapy, matrix
survival both on univariate and multivariate analyses. This metalloproteinases inhibitors or i.p. immunotherapy. There are
finding was similar to that from previous studies [5, 6, 8]. emerging new concepts in the treatment of malignant ascites
Patients with GI cancer were, however, found to be more likely on the basis of better understanding of the pathophysiological
to have liver metastases followed by breast cancer unlike our basis of its formation [14–17]. These new agents include i.p.
previous findings [5] where patients with breast cancer had the administration of a monoclonal antibodies (catumaxumab)
highest proportion of liver metastases a decade ago. Patients against epithelial cell adhesion molecule which has shown
with liver metastases in this study tended to be given diuretics encouraging results in a phase I/II trial [16] and is
on the understanding that their ascites is likely caused by the currently being investigated in a large multicenter randomised
renin–angiotensin–aldosterone pathway. Diuretics have been clinical trial.
shown to be more effective in the presence of liver metastases
[2, 10, 11].
Low levels of serum albumin and total proteins are significant conclusion
factors affecting survival adversely. In fact, low serum albumin is
The overall prognosis of patients with malignant ascites is poor
an independent prognostic factor especially in the nonovarian
and patients are typically in the palliative phase of care when
cancer groups. There are no other studies that have reported

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 16, 2014

they present. Patients, however, may be successfully palliated
the prognostic roles of serum albumin and total protein in
by careful selection and appropriate offer of therapy tailored
malignant ascites except our previous study [5]. The poor
towards individual patient’s characteristics and primary cancer
prognostic role is likely to be related to poor nutritional reserve,
type. The prognosis is, however, determined mainly by the
association of low total serum protein with the presence of liver
origin of the primary cancer, presence of hepatic metastases and
metastases and impaired immune network function in this
serum albumin concentration. Malignant ascites of ovarian
group of patients. Our findings also demonstrated that female
origin has a better prognosis even in the presence of ascites while
gender is associated with longer survival and better prognosis in
patients with malignant ascites of GI origin or unknown origin
malignant ascites and we believe that this is due to the combined
have the worst outcome.
preponderance of ovarian and breast cancer groups. Patients
Fifty-four per cent of our patients presented with ascites at the
who had chemotherapy also survived longer than those who
initial diagnosis of their cancer. The independent prognostic
did not receive chemotherapy but again this is likely to be
factors mentioned above should be used to select patients for
multifactorial such as response of cancer and ascites to
multimodality therapy as this may offer adequate palliation,
chemotherapy, the higher proportion of the ovarian cancer in
improved QoL and prolonged survival.
the chemotherapy group and the fact that patients selected for
chemotherapy are those with relatively good performance status
(PS) and longer life expectancy. We did not look at PS in this
retrospective study but we feel it is another important tool in references
the selection of patients suitable for aggressive treatment. PS is 1. Runyon BA. Care of patients with ascites. N Engl J Med 1994; 330: 337–342.
being used as part of the selection criteria for a multicentre 2. Parsons SL, Watson SA, Steele RJC. Malignant ascites [Review]. Br J Surg 1996;
prospective randomised therapeutic trial of the monoclonal 83: 6–4.
antibody catumaxumab and we await the results of this trial 3. Runyon BA, Hoefs JC, Morgan TR. Ascitic fluid analysis in malignancy-related
with interest. The patients who had no chemotherapy (50.7%) ascites. Hepatology 1988; 8: 1104–1109.
in this series were entered into the palliative care pathway and 4. Garrison RN, Kaelin LD, Galloway RH, Heuser LS. Malignant ascites: clinical and
had only symptomatic control of their malignant disease experimental observations. Ann Surg 1986; 203: 644–651.
5. Parsons SL, Lang MW, Steele RJC. Malignant ascites: a 2 year review from
process.
a teaching hospital. Eur J Surg Oncol 1996; 22: 237–239.
Serial paracentesis was still the commonest treatment offered
6. Mackey JR, Venner PM. Malignant ascites: demographics, therapeutic efficacy
to patients with symptomatic malignant ascites followed by the and predictors of survival. Can J Oncol 1996; 6: 474–480.
use of diuretics. Paracentesis provides relief of symptoms for 7. Wilailak S, Linasmita V, Srivannaboon S. Malignant ascites in female patients:
patients with malignant ascites and improves their QoL [12]. a seven-year review. J Med Assoc Thai 1999; 82: 15–19.
Lee et al. [13] in a survey of Canadian physicians’ practice 8. Appelqvist P, Silvo J, Salmela L, Kostiainen S. On the treatment and prognosis
showed that the use of diuretics was the second preferred of malignant ascites: is the survival time determined when the abdominal
treatment after therapeutic paracentesis, though it was not felt paracentesis is needed? J Surg Oncol 1982; 20: 238–242.
to be effective in the treatment of malignant ascites. The group 9. Ringenberg QS, Doll DC, Loy TS, Yarbro JW. Malignant ascites of unknown
of patients who had drainage at surgery and required no further origin. Cancer 1989; 64: 753–755.
10. Greenway B, Johnson PJ, Williams R. Control of malignant ascites with
drainage, and those who required no drainage at all during the
spironolactone. Br J Surg 1982; 69: 441–442.
course of their hospital episode, were those with subclinical
11. Pockros PJ, Esrason KT, Nguyen C et al. Mobilization of malignant ascites with
ascites that were either diagnosed radiologically or incidentally diuretics is dependent on ascitic fluid characteristics. Gastroeneterology 1992;
found at laparoscopy or laparotomy. There is no prospective 103: 1302–1306.
study evaluating the role of diuretics and no general consensus 12. Easson A, Bezjak A, Ross A et al. Changes in symptoms after paracentesis for
for the use of diuretics in the treatment of malignant ascites. symptomatic malignant ascites. J Clin Oncol 2005; 16: 6071 (Suppl).

948 | Ayantunde & Parsons Volume 18 | No. 5 | May 2007

Annals of Oncology original article
13. Lee CW, Bociek G, Fraught W. A survey of practice in management of malignant 16. Jaeger M, Stroehlein MA, Schoberth A et al. Immunotherapy with the trifunctional
ascites. J Pain Symptom Manage 1998; 16: 96–101. antibody removab leads to significant elimination of tumour cells from malignant
14. Marme A, Straub G, Bastert G et al. Intraperitoneal bispecific antibody ascites in ovarian cancer: results of a phase I/II study. J Clin Oncol 2004; 22
(HAE125XOKT3) therapy inhibits malignant ascites production in advanced (4S): 2504.
ovarian carcinoma. Int J Cancer 2002; 101: 183–189. 17. Byrne AT, Ross L, Holash J et al. Vascular endothelial growth factor-Trap
15. Heiss MM, Strohlein MA, Jaeger M et al. Immunotherapy of malignant ascites decreases tumour burden, inhibits ascites and causes dramatic vascular
with trifunctional antibodies. Int J Cancer 2005; 117: 435–443. remodelling in an ovarian cancer model. Clin Cancer Res 2003; 9: 5721–5728.

Downloaded from http://annonc.oxfordjournals.org/ by guest on January 16, 2014

Volume 18 | No. 5 | May 2007 doi:10.1093/annonc/mdl499 | 949