Abstract
Background:
Inconsistent results have been reported regarding mortality predictors of malignant pleural effusion (MPE).
Objective:
We aim to describe prognostic features of MPE in patients evaluated over 5 years in a referral center.
Methods:
All consecutive MPEs (confirmed by pleural cytology or biopsy) from January 2009 to July 2014 were included in this retrospective cohort study. Clinical data were collected from medical registries. Overall survival (OS) rate was calculated with Kaplan-Meier method and compared by the log-rank test. Hazard ratios (HRs) were calculated by the Cox proportional regression model.
Results:
A total of 110 patients were identified. Mean OS was 8.3 months (range: 25-426 days). No difference in OS was noted based on tumor type. Patients with MPE <50% had a significant better survival than those with malignant pleural effusion ≥50% (median OS: 15.7 vs 5.7 months; HR: 1.71; 95% confidence interval [CI]: 1.06-2.78;
Conclusion:
Our study showed that MPE size, as determined by chest X-ray, is significantly associated with mortality.
Introduction
Malignant pleural effusion (MPE) develops when tumor cells directly infiltrate the pleura and denoted a reduced life expectancy for patients with cancer. 1 Malignant pleural effusion represents advanced malignant disease, with median survival ranging from 3 to 12 months depending on the type of the underlying malignancy. 2 For these patients, some palliative treatment options can be offered based on tumor type, functional status, pleural apposition, and expected survival. 3,4 Nevertheless, a better identification of patients with the worst outcomes is mandatory in order to guide their medical management. Multiple studies have attempted to define mortality predictors for MPE based on clinical characteristics and biochemical parameters; however, the results have been inconsistent. 1,5 -7 Therefore, we aim to describe clinical, biochemical, and radiological prognostic features of MPE in consecutive patients evaluated over 5 years in our referral center.
Materials and Methods
We retrospectively reviewed a cohort of consecutive patients with MPE who were treated in our referral center (Hospital Mexico, San José, Costa Rica). We identified cases from an institutional database from January 2009 to July 2014. Inclusion criteria included age more than 18 years and confirmed MPE by pleural cytology or biopsy. We excluded cases with missing data on any of the variables we consider in the hypothesized models. The study was reviewed and approved by the institutional review board (CLOBI Hospital México 2012-22), and all patients gave informed consent to the work. Clinical data, radiological features, and biochemical analyses were collected from medical records. Analyzed variables included age; gender; presence of previous malignancy; symptoms including dyspnea, chest pain, or self-reported weight loss during the last 3 months; performance status (determined by Karnofsky scale); pleural fluid analyses including glucose, lactate dehydrogenase (LDH), and pH; cytological and histological studies as well as the effusion size determined by the radiologist on charge based on conventional chest X-ray (<25%, 25%-50%, 50%-75%, and >75% of the compromised lung).
Statistical Analysis
Continuous variables are presented as mean ± standard deviation. Categorical variables are presented as percentages. Survival time was determined from the day of MPE diagnosis to the day of death. Cases were censored at August 2014 or death. Overall survival (OS) rate was calculated with the Kaplan-Meier method and compared by the log-rank test. Hazard ratios (HRs) were calculated by the Cox proportional hazard regression model. We preselected potential prognostic variables based on previous studies.
2,8
Statistical significance was established at
Results
During follow-up time, we identified 110 patients with MPE. The diagnosis was confirmed by positive cytology in 70 patients and by biopsy in 40 patients. Table 1 summarizes the clinical characteristics of the entire cohort. Median follow-up time was 33.3 months and median overall survival (OS) was 8.3 months (range: 25-426 days; Figure 1). Using the Kaplan-Meier method, no difference in OS based on tumor type was noted (
Overall survival probability (Kaplan-Meier Method) of the studied population.
Clinical Characteristics of the Study Population.
Abbreviation: CT, computed tomography.
Patients with MPE < 50% had a significant better survival than those with MPE > 50%, with a median OS of 15.7 months compared to 5.7 months (HR: 1.71; 95% CI: 1.06-2.78;
Overall survival probability according to malignant pleural effusion size (HR: 1.71 [95% CI: 1.06-2.78]; log-rank test
No predictive prognostic value on mortality was identified by quartile distribution of age, pH, or glucose level in pleural fluid. Patients with pH ≤ 7.30 and glucose ≤60 mg/dL were compared with those who had pH > 7.30 and glucose > 60 mg/dL, and no significant survival differences were found (
Multivariate Analysis of Potential Prognostic Mortality Factors.
Abbreviation: KPS, Karnofsky Performance Status.
aStatistically significant
Discussion
Determinants of OS in patients with MPE are useful to stratify them according to their prognosis and guide their palliative management. In this study, we demonstrated that MPE size, as determined by chest X-ray, is an independent prognostic variable for mortality in patients with this condition. This cheap and very available technique can identify patients with the poorest prognosis and thus minimize unnecessary procedures in their final stages of life. It has been demonstrated that patients with a poor prognosis may prefer a therapeutic pleural aspiration to pleurodesis in order to relieve their symptoms. 8
A number of studies have reported risk factors for survival in patients with MPE. Unfortunately, some of these results are inconsistent and few of them have been carried out in developing countries where the access to some diagnostic methods and procedures are sometimes unreachable. Furthermore, ethnic and epidemiological factors can vary from industrialized countries where the majority of trials have been implemented.
Some authors have reported contrasting results about the relationship between massive pleural metastatic involvement and mortality. For example, Rodriguez-Panadero and colleagues 9 investigated the correlation between OS and the extent of tumor spread by direct pleural visualization by thoracoscopy, but no correlation with pleural effusion size was reported. On the other hand, the report of Jiménez and colleagues showed a direct association between massive MPE and mortality. 10 Our findings are in accordance with this latter report, since we found an increased risk of mortality for each quartile increase in MPE size. Thus, massive and nonmassive MPEs are independently associated with poor survival after adjusting for confounding variables.
In contrast with previous studies, 6,9 -12 we did not find any significant association between biochemical parameters in pleural fluid (ie, low glucose, high LDH, and low pH) and overall mortality. These discrepancies can be attributable to sample handling or as a consequence of a different composition of the cohorts. For example, the study of Ozyurtkan and colleagues 13 contained almost one-quarter of patients with mesothelioma, and the report of Rodriguez-Panadero and Lopez-Mejias 9 included 17% of patients with this diagnosis, as opposed to our cohort, which did not include any patient with this malignant condition. Besides, the relationship between pH and survival has a marginal prognostic value for estimating mortality in patients with MPE, 14 and its association with poor outcomes is currently under debate. The same is true for pleural fluid glucose, which has been inconsistently related to OS in previous trials. 15
Our population was mainly composed of patients with lung and breast cancer (67.3%), as similarly reported by previous series. 1,3 -10 Although we didn’t find any statistical difference in survival based on tumor type, we observed a shortest OS in patients with lung cancer compared to those with breast cancer. Some authors have shown this same difference in survival according to tumor type. 16 Nevertheless, median OS for women with breast cancer was better in our cohort than in the study of Clive and colleagues 17 (395 vs 192 days) as well as in patients with lung cancer (249 vs 74 days).
There are several limitations to our study including the retrospective design and the small number of patients. Besides, the unicenter study design limits its external validity. Nevertheless, giving the heterogeneity of data regarding this subject, we consider our findings valuable in order to better identify predictors of OS in this particular scenario. Further studies are necessary to guide clinicians for treatment of these patients.
Conclusion
To the best of our knowledge, our study provides the first documentation that MPE size as determined by chest X-ray is independently associated with survival. Thereby, chest X-ray can be used as a predicting tool in patients with MPE, particularly in developing countries where the pleural ultrasound or computed tomography scan is not readily available.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.