Article Text


A retrospective review of the palliative surgical management of malignant pleural effusions
  1. David Bell and
  2. Gavin Wright
  1. Department of Cardiothoracic Surgery, St Vincent's Hospital, Melbourne, Victoria, Australia
  1. Correspondence to Dr David Bell, Department of Cardiothoracics, St Vincent's Hospital Melbourne, 41 Victoria Parade, Fitzroy, Melbourne, VIC 3065, Australia; dovib{at}


Background A malignant pleural effusion (MPE) can be associated with debilitating dyspnoea. Determining optimal surgical management involves balancing the quality of palliation, degree of short-term dysfunction and expected duration of patient survival.

Objective To examine differences in postoperative survival after different surgical approaches to MPEs and to analyse the relationship between primary tumour site and duration of postoperative survival.

Design, setting and subjects Data from patients who underwent implantation of a tunnelled pleural catheter, video-assisted thoracoscopic surgery (VATS) pleurodesis or decortication at a thoracic surgical service from February 2001 to October 2010 were retrospectively reviewed.

Measurements and results 291 procedures were performed in 283 patients. The three most common primary malignancies were non-small cell lung cancer (NSCLC) (n=69), breast (n=62) and mesothelioma (n=49). Median postoperative survival was 184 days for NSCLC patients, 221 days for breast cancer patients and 595 days for mesothelioma patients. Tunnelled pleural catheters were implanted in 33 patients with a median survival of 92 days. VATS pleurodeses were performed in 224 patients with a median survival of 227 days. Decortications were performed in 26 patients with a median survival of 379 days.

Conclusions Results reflect that procedures can be successfully matched to projected survival, although prognosis is not informed by primary tumour site, age or sex. Further to this, results are consistent with the policy that the procedures resulting in the shortest hospital stay and lowest peri-operative morbidity should be reserved for patients with the shortest predicted survival.

  • Pleural effusion, malignant
  • Pleurodesis
  • Non small cell lung cancer
  • Thoracic surgery, Video-assisted

Statistics from


A malignant pleural effusion (MPE) is defined by the presence of malignant cells in the pleural fluid or the presence of a pleural effusion in the setting of a pleural biopsy that indicates malignancy. An MPE suggests disseminated malignancy and is associated with a poor prognosis.1

The incidence of MPEs is 660 per million per year in Australia.2 While up to 25% of patients with MPEs are asymptomatic, most present with dyspnoea considerably affecting quality of life. Other possible symptoms include orthopnoea, cough and chest discomfort.3 Although the eradication of an MPE does not necessarily improve patient survival, in most cases it can control symptoms and improve quality of life.4

The challenge facing clinicians is to recommend a palliative surgical procedure to patients with MPEs based on their predicted survival course. That is to say, while a pleurodesis or decortication may provide more definitive symptom relief, subjecting patients with a very limited prognosis to a long hospital stay and possible peri-operative complications may be counterproductive. This study aims to examine the differences in patient survival and length of hospital stay after three different palliative surgical approaches for MPEs: implantation of a tunnelled pleural catheter, video-assisted thoracoscopic surgery (VATS) pleurodesis and decortication. This study also aims to analyse the relationship between primary tumour site and duration of postoperative survival.


Patients who underwent implantation of a tunnelled pleural catheter, VATS pleurodesis or decortication at a combined thoracic surgical oncology practice at St Vincent's Hospital Melbourne, St Vincent's Private Hospital and Peter MacCallum Cancer Centre from 13 February 2001 to 10 October 2010 were included in this study.

The PleurX catheter system (CareFusion Corporation, California, USA) was used as the tunnelled pleural catheter of choice. PleurX catheters were implanted in patients with symptomatic MPEs in whom other procedures had failed or a general anaesthetic would not be tolerated or was not desired. VATS pleurodeses were performed in patients with recurrent symptomatic MPEs and a predicted survival of more than 3 months. In all cases pleurodesis was attempted by VATS with talc insufflation. Decortication was performed in patients with a predicted life expectancy of more than 6 months who also had considerable lung entrapment that precluded pleurodesis if there was a likelihood of a dissection plane between the entrapping rind and the underlying visceral pleura.

The hospital records of all patients were retrospectively analysed, and the following data were recorded: date of birth, sex, site of primary malignancy, procedure performed, duration of hospital stay (calculated from the date of the procedure to the date of discharge) and date of death. If no death was recorded in the hospital database, the patient's most recent outpatient appointment or hospital visit was used as the last time the patient was noted to be alive with disease.


During the study period, 291 procedures were performed on 283 patients. Deaths were recorded for 181 patients in the hospital databases. Patients for whom no date of death was recorded were excluded from survival analysis.

The three most common primary malignancies in this series were non-small cell lung cancer (NSCLC) (n=69), breast cancer (n=62) and mesothelioma (n=49). Other common primary tumours included sarcoma (n=16), melanoma (n=12) and colorectal cancer (n=9).

Table 1 provides a summary of the age and sex breakdown of the patients included in the analysis. Analysis of variance shows that men were significantly older than women (p=0.014). There was no significant difference between ages of patients undergoing different procedures although patients with breast cancer were significantly younger than those with either NSCLC or mesothelioma (p=0.003). There was not a significant difference in survival duration (averaged over procedures and malignancies) between men and women.

Table 1

Age details of the sample

During the study period, 33 PleurX catheters were implanted. Three were implanted after failed pleurodeses and one after a failed decortication. The most common primary tumours in this group were NSCLC (n=14), breast cancer (n=6) and melanoma (n=4). The patients with implanted PleurX catheters had the shortest median duration of hospital stay (2 days). VATS pleurodeses were performed on 229 occasions, and in the case of five patients, were performed bilaterally and sequentially. The most common malignancies in this group were breast cancer (n=52), NSCLC (n=53) and mesothelioma (n=28). The median duration of postoperative hospital stay for this group was 7 days. A total of 27 decortications (17 by VATS and 10 by thoracotomy) were performed in patients with considerable lung entrapment. Two decortications were carried out after failed attempts at pleurodesis, and one decortication was repeated after an initial failure of lung re-expansion. The primary tumour in the majority of these patients was mesothelioma (n=19). Patients who underwent decortication had the longest median duration of hospital stay (8 days).

Median survival times were 184 days (95% CI 132–437 days) for NSCLC patients, 221 days (136–416) for breast cancer patients and 595 days (379–783) for mesothelioma patients. It is clear that there was little difference between survival times for NSCLC and breast cancer patients, but mesothelioma patients tended to have longer survival times. A summary of median survival times and predicted 12-month survival is provided in table 2. Figure 1 represents the Kaplan–Meier curve for malignancy type.

Table 2

Median and 12-month survival times based on gender, procedure and malignancy type

Figure 1

Kaplan–Meier survival curves for malignancy type. NSCLC, non-small cell lung cancer.

A Cox proportional hazards model was fitted to the data, including age and gender as predictors, and showed that neither of these factors had a significant effect on the results. The results are summarised in table 3. The relative risks can be interpreted as follows: for age, a 1-year increase in age increases the relative risk of death by 1.1%, while men show a 5.9% increase in relative risk of death over women. For malignancy type, breast cancer shows a reduced but non-significant relative risk of death relative compared with NSCLC, the reference malignancy. For mesothelioma, the relative risk of death is only 45% of the risk of death from NSCLC.

Table 3

Details of fit for Cox model for malignancy type

Median survival time for patients postdecortication was 379 days (95% CI 246 days–∞) and for patients postpleurodesis was 227 days (189–332 days). Survival time for patients post-pleurX catheter implantation was significantly shorter at 92 days (45–176 days). Figure 2 represents the Kaplan–Meier curve for procedure type.

Figure 2

Kaplan–Meier survival curves for procedure type.

Details of the Cox model fitted to these data are presented in table 4. The Cox model shows a significant effect of pleurX catheter (with decortication as the reference level), but not age or gender. The relative risk of death for pleurX (relative to decortication) is 3.95, adjusted for age and gender.

Table 4

Details of fit for Cox model for procedure type


The approach to the management of an MPE is a difficult problem in palliative care practice. Management is guided by an estimation of the patient's prognosis and a balance of the likely morbidity and benefit of the proposed procedure. This can often be impossible to predict.

Once an MPE is confirmed, management options include observation, implantation of a tunnelled pleural catheter, pleurodesis or a decortication. Observation is indicated if the pleural effusion reaccumulates very slowly after an initial diagnostic thoracentesis or if drainage does not result in significant symptom relief owing to patient comorbidities. Conservative management may also be indicated if a patient's predicted survival is very poor.

Implantation of a tunnelled pleural catheter involves placement of a silicone catheter into the pleural space and subsequent outpatient drainage via vacuum bottles. This procedure is usually considered for patients who are deemed unsuitable to undergo pleurodesis by VATS or who choose a completely outpatient-based therapy.

The gold standard procedure for the management of an MPE is pleurodesis. This involves eradication of the pleural space by inducing inflammation and subsequent scarring between the parietal and visceral pleura. This can be performed by VATS with insufflation of a sclerosant and/or mechanical abrasion. Alternatively, it can be performed under local anaesthesia at the bedside by retrograde injection of talc slurry into a chest tube.

Decortication is an invasive procedure involving removal of the organising fibrinous rind from the visceral pleura. It may be performed by thoracotomy or VATS and is indicated in patients with considerable lung entrapment precluding pleurodesis. Lung entrapment may be caused by either a high burden of pleural disease, as in the case of mesothelioma, or activation of coagulating proteins in the pleural exudates.

MPEs are most commonly caused by a primary lung or breast cancer. Data compiled from five recent studies revealed that the primary tumour sites for MPEs are lung (37.5%), breast (16.8%), lymphoma (11.5%), the genitourinary tract (9.4%), the gastrointestinal tract (6.9%) and unknown (10.7%).1 Studies from the 1980s indicate that the incidence of MPEs caused by mesothelioma is approximately 4%.1 However, there is no doubt that the incidence has increased since then as rates of mesothelioma related to asbestos exposure in the mid-twentieth century are approaching their peak.5

The distribution of primary tumour type in this series was different to that reported in most previous studies with higher rates of breast cancer (21.9%) and mesothelioma (17.3%). The higher rates of mesothelioma reflect the fact that Australia has one of the highest rates of mesothelioma worldwide at 30 per million per year. This is in contrast to rates of nine per million per year in the USA and similar rates in Western Europe where the majority of previous studies relating to the management of MPEs were conducted.6

The median survival after the diagnosis of an MPE varies widely from 3 to 12 months in the literature, depending on the primary tumour site. Historically, studies have shown that patients with malignant effusions secondary to lung cancer have the shortest survival time and patients with malignant effusions secondary to breast cancer tend to have a longer survival course. Heffner et al reported that the median survival of 60 patients with malignant effusions secondary to breast cancer was 5 months, while the median survival of 146 patients with primary lung cancer was 3 months.7 Results from this study compare favourably, with breast and lung cancer patients having a similar median postoperative survival of 7 and 6 months, respectively. The 12-month survival for patients with NSCLC (38.2%) and breast cancer (36.6%) in this series was also similar indicating that the site of primary malignancy alone is not a fair indicator of prognosis and should not guide surgical treatment.

Evaluation of a patient's prognosis is a difficult step in determining how to best manage an MPE. While there was no formal attempt in this study to find a relationship between patient factors and prognosis, Cox proportional hazard models showed that for both different malignancies and procedures neither age nor gender had had a significant impact on postoperative survival time. The postprocedure survival was consistent with desired outcomes. That is to say patients who underwent more invasive procedures such as a decortication or VATS pleurodesis generally outlived those who underwent the less invasive PleurX catheter implantation. This indicates that although no specific algorithm was used, the evaluation of patient prognosis in this series was accurate.

In addition to considering prognosis, the decision to perform PleurX catheter implantation, VATS pleurodesis or decortication involves balancing the morbidity of the proposed procedure with the potential improvement in the quality of life.

In this study, the length of postoperative hospital stay was used as a surrogate marker of postoperative morbidity. As expected, the level of invasiveness of the procedure, as determined by the predicted patient survival, reflected the required duration of hospital stay. Patients who underwent implantation of a PleurX catheter had the shortest length of hospital stay and those who underwent a decortication had the greatest median length of hospital stay.

The obvious disadvantage of a PleurX catheter is the constant discomfort and psychological reminder associated with a permanent chest tube. The main advantage is that it does not require administration of a general anaesthetic or an extended hospital stay. In addition, the procedure-related mortality and morbidity are minimal. Despite the simplicity of the procedure, the PleurX catheter has been shown to effectively relieve dyspnoea related to MPEs with minimal complications.8

Multiple studies have shown that VATS pleurodesis is an efficacious and relatively safe procedure. In a study by Cardillo et al spanning 9 years, 611 patients with MPEs were treated with VATS pleurodeses. The operative mortality was 0.81% (five cases) and the postoperative complications (19 cases) were relatively minimal and mostly related to prolonged air leak. Long term follow-up showed that 92.7% of patients had no effusion recurrence.9 Other authors reported similar results, with Schulze et al10 reporting a 92.2% effusion non-recurrence rate in a series of 105 VATS pleurodeses.

Despite the fact that VATS pleurodesis is rather safe, it is an invasive procedure that requires general anaesthesia. Many centres therefore arbitrarily recommend the procedure to patients with a predicted life expectancy of more than 3 months. Complications of this procedure are not uncommon but are usually benign relating to prolonged air leak (usually a persistent space from lung entrapment). It is important to warn patients of less common side effects that may affect the remaining quality of life including considerable postoperative neuralgia and pleural space or wound infection.4

Open decortication requires a good performance status and a predicted survival of at least 6 months to justify the significant morbidity and risk of mortality associated with the procedure.1 Of the 27 decortications performed in this study, 19 were performed in patients with mesothelioma. This reflects the fact that mesothelioma is a pleurally-based malignancy with a tendency towards forming a tumour and fibrous rind that causes lung entrapment. When compared with patients from other tumour streams, those with mesothelioma had the longest duration of postprocedure survival, partly because they collectively had less overt disease outside the thorax. This justifies the use of a more invasive but definitive procedure.

There are obvious limitations to this study given that this is a retrospective analysis compiled from case notes. Data reflecting different patient characteristics such as comorbidities and performance status were not collected, although age and gender were shown not to have a significant survival effect. Data reflecting patient comorbidities and performance status may have helped outline variables that influence outcomes in these patients. As well as this, 102 patients (36%) had no recorded date of death, limiting the number of patients able to be used for survival analysis.

The findings in this study provide insight into the practice at a busy thoracic oncology service over an extended period of time. Results reflect that procedures can be successfully matched to projected survival, although prognosis is not informed by primary tumour site, age or sex. Further to this, results are consistent with the policy that the procedures resulting in the shortest hospital stay and lowest peri-operative morbidity should be reserved for patients with the shortest predicted survival.


We would like to acknowledge Mr John Field from John Field Consulting who provided an indepth statistical analysis of our results.


View Abstract


  • Contributors DB carried out the data collection and drafted the manuscript. GW was responsible for editing and approving the article.

  • Competing interests None.

  • Ethics approval This study was approved by the Human Research Ethics Committee of St Vincent's Hospital, Melbourne.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The data from this study are available with the primary author, David Bell, and coauthor Gavin Wright. There are no additional data that have not been presented which are available.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.