Ann Surg Oncol DOI 10.1245/s10434-015-4491-6
ORIGINAL ARTICLE – THORACIC ONCOLOGY
Prognostic Factors of 30-Day Mortality After Palliative Procedures in Patients with Malignant Pleural Effusion Fernando Conrado Abrao, MD1, Igor Renato Louro Bruno de Abreu, MD2, Mariana Fogarolli, MD2, Giovanna Caxeiro, MD2, Camila Borges Saes Bezerra, MD2, Fernanda Prado de Cerqueira Cesar, MD2, Poline Spitti Rocha, MD2, and Riad Naim Younes, PhD1 Department of Thoracic Surgery, Hospital Sa˜o Jose´, Sa˜o Paulo, SP, Brazil; 2Department of Thoracic Surgery, Hospital Santa Marcelina, Sa˜o Paulo, Brazil
1
ABSTRACT Background. The aim of this study was to identify predictors of 30-day mortality in patients with malignant pleural effusion (MPE) who need pleural palliative procedures. Methods. Data was prospectively collected from our database, between January 2013 and July 2014 on 86 patients with MPE and complete follow-up 30 days after the procedure. The sample was divided into two groups. The first group (G1) included patients who had died up to 30 days after the palliative procedure, and the second group (G2) included patients who survived for more than 30 days after the palliative procedure. The identification of prognostic factors occurred through univariate analysis using Fisher exact test for analysis of categorical variables and the t test for quantitative variables. Subsequently, all variables were used in the multivariate logistic regression analysis. The cutoff values for any significant continuous variables were determined by receiver operating characteristics analysis. Results. There were 24 patients in G1 and 62 patients in G2. Univariate analysis of factors affecting postprocedural survival disclosed nine factors that were associated with significantly reduced postoperative survival. At the multivariate analysis, high levels of white blood cells, (p = .013), low levels of red blood cells (p \ .0001) and protein in pleural fluid (p = .001), and primary lung and gastrointestinal sites (p = .0076) were identified as independent predictors of mortality.
Ó Society of Surgical Oncology 2015 First Received: 30 January 2015 F. C. Abrao, MD e-mail: [email protected]
Conclusions. We identified four factors that are easily recognized in daily practice and can help select patients with low life expectancy. Therefore, invasive procedures and hospitalizations for this subgroup of patients can be prevented.
Malignant pleural effusion (MPE) is common, and given the year-on-year increase in new cancer diagnoses the incidence is set to rise.1,2 MPE represents advanced malignant disease, and current guidelines report median survival rates \1 year and decreased quality of life.3,4 Pleural and oncological treatment options are increasing, and hence more accurate prognosis at presentation may help to identify patients with the poorest prognosis and thus minimize unnecessary procedures and discomfort in their final stages of life.5,6 A few studies have shown predictors of survival in patients with malignant pleural disease, including tumor characteristics, extent of disease, comorbidities, and the composition of the effusion, in addition to inflammationbased scoring systems.7–12 However, to date, no study has attempted to identify predictors of mortality at 30 days, which would greatly help to choose less-invasive palliation options. The aim of this study was to identify predictors of 30-day mortality in patients with malignant pleural effusion who need pleural palliative procedures. MATERIALS AND METHODS A prospectively collected database from our institution, between January 2013 and July 2014, was used to identify patients with MPE submitted to the first pleural palliative procedure. All patients had been followed for a minimum of 1 month or until death.
F. C. Abrao et al.
The ethics committee of the university hospital of Santa Marcelina medical school in Sa˜o Paulo, Brazil, approved this study, under protocol number 39462714.2. Malignant pleural effusion was defined as cases with positive malignant cells in pleural fluid, those with neoplastic pleural infiltration identified in anatomopathological assessment or in patients unfit for pleuroscopy and pleural biopsy, but with metastatic cancer at other sites confirmed by the pathological study, and pleural effusion with no other diagnosed cause after review of the clinical team and pleural fluid analysis. Data on characteristics available preoperatively were collected to allow analysis of prognostic factors, which may influence survival. Collected data included basic demographics, American Society of Anesthesiologists (ASA) physical status classification, performance status according Eastern Cooperative Oncology Group (ECOG) score, hematology parameters, including white blood cells (WBC), number of neutrophils and lymphocytes, neutrophil/lymphocyte ratio (NLR) red blood cells (RBC), and body mass index (BMI) on the day before surgery. The influence of the primary tumor site was also assessed. During the period after the palliative procedure, we evaluated the volume of drained liquid, the type of palliative procedure performed, recurrence of pleural effusion, and the presence of neoplastic cells in pleural fluid, in addition to the biochemical profile including pleural fluid pH, levels of adenosine deaminase (ADA), total protein, albumin, glucose, lactate dehydrogenase (DHL), and the proportion of lymphocytes in pleural fluid (Table 1). Referring physicians and oncologists were contacted to obtain survival data. The palliative method was defined after multidisciplinary discussion. Thoracocentesis was used in those patients who the clinical team thought would have pleural response to chemotherapy, and, therefore, more invasive procedures would not be required. The remaining patients underwent more invasive procedure. In patients considered unfit for general anesthesia, an intercostal drain was inserted under local anesthetic procedure. Once the lung had reexpanded and the drainage reduced, talc slurry was injected into the pleural cavity through the drain. Also in patients considered unfit for general anesthesia, we started using an indwelling pleural catheter in 2014, when it was made available at our institution. Patients treated under general anesthetic procedure underwent an initial bronchoscopy to exclude central airway obstruction. With the patient in the lateral decubitus position, a single port was fashioned to allow inspection of the pleural space with video-assisted thoracoscopy (VAT). After appropriate effusion samples were obtained for cytological examination and culture, multiple large pleural biopsies were taken from any abnormal areas and, if the pleura were uniform, from a number of representative
areas, including the posteroinferior costal pleura. If no preoperative diagnosis was available, these were submitted to frozen-section histopathological examination. If lung reinflation was deemed adequate to allow apposition of the pleural surfaces over most of the hemithorax, then pleurodesis was attempted by insufflation of 4 g of sterile talc. Where the pleural space was obliterated by the tumor, no procedure other than pleural biopsy was performed. In patients undergoing VAT or intercostal drain inserted under local anesthesia, those in whom lung expansion was not greater than half of the hemithorax had the tube removed after 24 h after the procedure and the patient was submitted to monthly revaluations to verify the necessity of new pleural procedure (thoracentesis or indwelling catheter). We divided our sample into two groups. The first group (group I) included patients who had died up to 30 days after the palliative procedure, and the second group (group II) included patients who survived more than 30 days after the palliative procedure. The identification of prognostic factors was performed through univariate analysis, using Fisher exact test for the analysis of categorical variables and t test for quantitative variables. Subsequently, all variables were used in the multivariate logistic regression analysis. This initial model was refined to originate the final model with the significant variables. The p values of .05 or less were considered significant. The cutoff points for any significant continuous variables were determined by receiver operating characteristics (ROC) analysis. Calculations were performed using the statistical software SPSS, version 20.0.
RESULTS A total of 94 consecutive patients underwent the first palliative procedure for MPE during an 18-month period. The most common causes were breast and lung carcinoma. All patient characteristics are shown in Table 1. Eight patients were excluded, as we were unable to get in touch with them 30 days after the pleural procedure, totaling 86 analyzed patients (Table 2). There were 24 patients in group I and 62 patients in group II. The causes of death in group I (perioperative mortality) were pulmonary embolus, in nine; respiratory failure, in 11; hepatic failure, in two; acute abdomen perforation, in two. None of the patients with leukocytosis had pleural empyema at the surgical procedure. The procedures performed were 23 bedside drainages (25.5 %), 20 bedside pleurodesis through a drain (24.5 %), 16 VAT pleurodesis (19.1 %), 17 thoracocenteses (19.2 %), 8 cases of indwelling pleural catheter (9.6 %), and 2 VAT pleural biopsies (2.1 %). Overall median postoperative survival was 127 days.
30-Day mortality After Palliative Procedures TABLE 1 Characteristics of the study population (n = 94) Characteristics
n
%
TABLE 2 Univariate analysis of possible factors related to survival at 30 days (n = 86) Characteristic/group
Gender Female
67
71.3
Male
27
28.7
No (% or SD)
Breast
32
Male
27.7 34.0
17 (27.9)
44 (72.1)
7 (28.0)
18 (72.0)
6 (27.3)
16 (72.7) 26 (83.9)
Primary site
.0463
Gastrointestinal
13
13.8
Lung
Othersa
23
24.2
Breast
5 (16.1)
Gastrointestinal
8 (61.5)
Othersa
5 (25)
ASA 1
1
1.1
2 3
37 49
39.4 52.1
4
7
7.4
Yes (% or SD) 1.0000
Female
26
p value
Gender
Primary site Lung
Survival [30 days
5 (38.5) 15 (75) \.0001
ASA 1
ECOG
0 (.0)
1 (100.0)
2
2 (6.2)
30 (93.8)
3
16 (34.8)
30 (65.2)
4
6 (85.7)
1 (14.3)
0
0 (.0)
4 (100.0)
0
4
4.3
1
27
28.7
2
33
35.1
1
0 (.0)
23 (100.0)
3
22
23.4
2
7 (22.6)
24 (77.4)
4
8
8.5
3
12 (60.0)
8 (40.0)
4
5 (62.5)
3 (37.5) 13 (56.5)
\.0001
ECOG
Procedure Pleurodesis with drainage
23
24.5
VAT pleurodesis
18
19.1
Drainage
10 (43.5)
Drainage
24
25.5
PleureX
4 (50.0)
4 (50.0)
Thoracocentesis
18
19.1
Pleurodesis with drainage
3 (15.0)
17 (85.0)
Vat pleural biopsy
2
2.1
Pleurodesis with vats
2 (12.5)
14 (87.5)
Indwelling catheter
9
9.6
Thoracocentesis
4 (23.5)
13 (76.5)
Vats
1 (50.0)
1 (50.0)
13 (28.3)
33 (71.7)
Pleural histology/cytology Negative Positive
Procedure
50
53.2
Histocytology
44
46.8
Negative
11 (27.5)
29 (72.5)
81.9
Age (years)
Yes
1.0000
59.6 (11.8)
59.6 (14.3)
.7689
18.1
BMI (kg/m2)
22.8 (4.5)
24.7 (4.1)
.2324
Volume (mL)
902.1 (389.7)
970.8 (482.1)
.4738
ADA (IU/L)
9.8 (6.5)
12.9 (10.3)
.2302 .3090
Positive
Recurrence No
.0887
77 17
ASA American Society of Anesthesiologists (ASA) physical status classification, ECOG Eastern Cooperative Oncology Group performance status score
pH
7.6 (.5)
7.7 (.4)
584.7 (646.4)
466.9 (355.0)
.5202
93.0 (79.3)
100.0 (36.3)
.1543
Others: renal, lymphoma, gynecological, kidney, prostate, and thyroid
LDH (IU/L) Albumin (g/dL)
1.9 (.7)
2.8 (.5)
Univariate analysis of factors affecting postprocedural survival verified that primary tumor site and hemoglobin levels, in addition to seven other variables (Table 2), were associated with a significantly reduced postoperative survival. At the multivariate analysis, high levels of WBC (p = .013), low levels of RBC (p \ .0001), and pleural fluid protein (p = .001), as well as lung and gastrointestinal primary site (p = .0076) were identified as independent predictors of mortality (Table 3). ROC analysis identified a WBC value of 9 9 109/L, RBC of 11 g/dL, and total protein of pleural fluid of 3.6 g/dL as the decision-making
Protein (g/dL)
3.3 (1.1)
4.3 (.9)
\.0001
Hemoglobin (g/dL)
10.1 (1.7)
12.0 (1.9)
\.0001
Leukocytes (109/L)
12229 (6100)
8934 (4717)
.0161 .0047
a
Glucose (g/dL)
9
Neutrophils (10 /L) Lymphocytes (109/L) NLR
\.0001
9965 (5770)
6426 (4226)
1172.8 (762.6)
1499.0 (810.3)
.0831
10.2 (7.5)
5.6 (4.8)
.0008
SD standard deviation, ASA American Society of Anesthesiologists (ASA) physical status classification, ECOG Eastern Cooperative Oncology Group performance status score, BMI body mass index, ADA adenosine deaminase, LDH lactate dehydrogenase, NLR neutrophil/lymphocyte ratio a
Others: renal, lymphoma, gynecological, kidney, prostate, and thyroid
F. C. Abrao et al. TABLE 3 Independent predictors of mortality at 30 days at the multivariate analysis Characteristic
Level 2
Odds ratio
Lower 95 %
Upper 95 %
p value
Lung
Breast
17.4239
2.3208
222.1541
Lung
Breast
11.1742
1.2500
167.7473
Lung
Gastrointestinal
1.4158
.1846
11.7608
Gastrointestinal
Breast
12.3070
1.6514
128.1258
Gastrointestinal
Others
7.8926
.9161
98.5765
Breast
Others
1.5593
.1455
15.9397
Protein (g/dL)
.3061
.1226
.6356
.0010
Hemoglobin (g/dL)
.4201
.2311
.6690
\.0001
Leukocytes (109/L)
1.0002
1.0000
1.0003
.0130
A
100
75 50 25 0 0
10
20
75 50 25 0
30
0
10
20
days ≤9000
≤ 9000 : median= – > 9000 : median= 44
30
days ≤11
>9000
Log Rank p= 0.018
>11
Log Rank p= 0.0002 ≤ 11 : median= 30
IC 95% (-; -)
> 11 : median= -
IC95% (0.3 ; 79.7)
IC 95% (11.9 ; 48.1) IC95% (-; -)
B
A
100 100
90
cumulative survival
cumulative survival
FIG. 2 a Kaplan–Meier survival curve comparing patients with total protein from pleural fluid B3.6 g or [3.6 g/dL. b Kaplan–Meier survival curve comparing patients with 0, 1, 2, 3, or 4 prognostic factors determined by multivariate analysis
.0076
B 100
cumulative survival
FIG. 1 a Kaplan–Meier survival curve comparing patients with a white cell count B9 9 109 or [9 9 109/L. b Kaplan–Meier survival curve comparing patients with a red blood cell count B11 or [11 g/dL
cumulative survival
Primary site
Level 1
75 50 25
80 70 60 50 40 30 20
0 0
10
20
days ≤3,6
>3,6
30
10 0 0
10
none
> 3.6 : median= -
two
20
three
30
four
Log Rank p= 0.0001
Log Rank p= 0.0001
≤ 3.6 : median= 28
one
days
IC 95% (17.3 ; 38.7) IC95% (-; -)
none factor :
median= - IC 95% (-; -)
one factor :
median= - IC 95% (-; -)
two factors :
median= 81 IC 95% (0.0 ; 175.8 )
three factors: median= 16 IC 95% (0.0 ; 30.5 ) four factors:
median= 9.5 IC 95% (-; -)
30-Day mortality After Palliative Procedures
threshold, with the highest diagnostic accuracy for predicting death (Figs. 1, 2a). The presence of none, some, or all of factors, which maintain independent significant associations with 30-day mortality after the palliative procedure, stratifies the patient population into five distinct groups. The median survival of patients with all poor prognostic factors is significantly lower than those with none (81 vs. 9.5 days; p \ .001) (Fig. 2b).
DISCUSSION Survival after MPE palliative procedure is important to determine which procedure is more appropriate to avoid exposing patients with low life expectancy to higher morbimortality. Therefore, the more invasive techniques should be reserved for patients with longer survival. The cutoff of mortality at 30 days was chosen because in our country, culturally it is easier to explain and justify less aggressive procedures in patients known to have a life expectancy of 30 days. Additionally, no studies have evaluated mortality in this particular period, to the best of our knowledge. Our study disclosed 4 factors associated with survival \30 days and are easily identifiable: primary pulmonary and gastrointestinal site, serum hemoglobin \11 g/dL, protein concentration in the pleural fluid \3.6 g/dL, and serum leukocyte levels [9 9 109/L. In addition, the combination of these factors also showed a significant impact on patient survival (Fig. 2b). Therefore, it would be reasonable to submit these patients to palliative procedure with multiple pleural aspirations or the use of an indwelling pleural catheter inserted under local anesthesia. In our study, the histology of the primary tumor was an independent prognostic factor, with breast being the histological type of better prognosis and lung, together with gastrointestinal cancer, the histological types of worse prognosis. The only study that evaluated early mortality (3month survival), with a study methodology similar to ours, showed that breast cancer was also associated with longer survival.13 However, studies that evaluated the primary site as an overall survival prognostic factor have shown controversial results. Anevlavis et al. reported ovarian cancer with median survival of 18 versus 15 months for breast cancer (p = .008) being an independent prognostic factor.14 On the other hand, Pilling et al. described malignant pleural mesothelioma and breast cancer as showing a trend toward increased survival when compared with other primary sites, although this was not statistically significant.7 A lower concentration of pleural fluid protein was associated with a lower survival, in our study. Bielsa et al. showed a mean survival of 2.2 months when the pleural fluid total protein value was than 3.85 g/dL, and this was
statistically significant in the multivariate analysis.8 In the ¨ zyurtkan et al., aimed to identify predictors of study by O mortality in patients with MPE at 3 months, the low concentration of protein was only significant in the univariate analysis (p \ .0001) and the cutoff was B3.7 g/dL.13 All these studies included patients who had received previous oncologic treatment; i.e., MPE was a sign of disease progression. On the other hand, Anevlavis et al. studied 90 patients who had received no systemic treatment for cancer and, therefore, had less advanced disease. In this sample, total protein concentration in the pleural fluid was a factor not related to patient survival.14 The explanation may be the advanced stage of cancer, which is strongly associated with the hypoproteinemia and hypoalbuminemia.8,15 Increased WBC and neutrophil levels and an increased N/L ratio were statistically significantly correlated with poor survival in our univariate analysis. Only WBC was an independent factor predicting survival. Conversely to our study, Anevlavis et al. found that N/L ratio, and not WBC, was a predictor of mortality in the multivariate analysis.14 Pilling et al. showed that a leukocyte level [12 9 109/L was an independent predictor of survival.7 These data suggest that systemic inflammation is also an important prognostic factor.12,14,16 This is consistent with growing evidence that systemic inflammation may contribute to both morbidity (in the form of constitutional symptoms) and mortality in cancer and may represent a common pathway in the fatal progression of malignancy.11 Further studies are necessary to understand the mechanisms behind its prognostic value. As for anemia, few studies have associated it with survival in MPE. Pilling et al. showed that hemoglobin \11 g/dL is associated with lower overall survival in patients with MPE, but without statistical significance.7 However, in our study, there was an impact on 30-day survival. These findings seem to support the association between anemia and advanced disease. Performance status has been shown to be associated with prognosis in MPE patients.8,9,14,17 In our study, ECOG was a factor associated with 30-day survival in the univariate analysis, only. Perhaps if our cutoff survival had ¨ zyurtkan et al., we been 3 months as in the study by O would have obtained ECOG as a factor associated with survival, as they reported.13 There is no doubt that survival in patients with high ECOG is lower; however, the median survival of these patients varies, and it is greater than 30 days in the literature, with a 45-day median for ECOG 3 and 33 days for ECOG 3–4.14,18 Perhaps, as an isolated factor, ECOG is not a good parameter for predicting mortality at 30 days. Our study has some limitations. The population we have studied is the one referred to the thoracic surgical unit for consideration of palliation, and it is not disease-specific. We did not evaluate the oncologic treatment of these
F. C. Abrao et al.
patients; thus, we do not know its influence on survival. The group of patients who died at 30 days is much smaller (24 patients) than the group of 62 patients who died after 30 days. However, this disproportion is also found in the literature, in which Dresler et al. reported 17 % of mortality at 30 days.19 We conclude that the primary site, anemia, leukocytosis, and low protein concentration in pleural fluid are associated with 30-day mortality in patients with MPE submitted to the first pleural palliative procedure. The analysis of these factors can identify patients with low survival and, consequently, help physicians choose the most appropriate palliative procedures. DISCLOSURE
The authors declare no conflicts of interest.
REFERENCES 1. The American Thoracic Society. Management of malignant pleural effusions. Am J Respir Crit Care Med. 2000;162:1987–2001. 2. Cancer Statistics Registrations, England (Series MB1): Office of National Statistics, Stationary Office, 2010. 3. Antunes G, Neville E, Duffy J, Ali N; Pleural Diseases Group, Standards of Care Committee, British Thoracic Society. BTS guidelines for the management of malignant pleural effusions. Thorax. 2003;58(suppl II):29–38. 4. Antony VB, Loddenkemper R, Astoul P, et al. Management of malignant pleural effusions. Eur Respir J. 2001;18:402–19. 5. Roberts ME, Neville E, Berrisford RG, Antunes G, Ali NJ; BTS Pleural Disease Guideline Group. Management of a malignant pleural effusion: British Thoracic Society Pleural Disease Guideline 2010. Thorax. 2010;65(Suppl 2):32–40. 6. Tan C, Sedrakyan A, Browne J, Swift S, Treasure T. The evidence on the effectiveness of management for malignant pleural effusion: a systematic review. Eur J Cardiothorac Surg. 2006;29:829–38. 7. Pilling JE, Dusmet ME, Ladas G, Goldstraw P. Prognostic factors for survival after surgical palliation of malignant pleural effusion. J Thorac Oncol. 2010;5:1544–50.
8. Bielsa S, Salud A, Martinez M, et al. Prognostic significance of pleural fluid data in patients with malignant effusion. Eur J Intern Med. 2008;19:334–9. 9. Burrows CM, Mathews WC, Colt HG. Predicting survival in patients with recurrent symptomatic malignant pleural effusions: an assessment of the prognostic values of physiologic, morphologic, and quality of life measures of extent of disease. Chest. 2000;117:73–8. 10. Pinato DJ, Mauri FA, Ramakrishnan R, Wahab L, Lloyd T, Sharma R. Inflammation-based prognostic indices in malignant pleural mesothelioma. J Thorac Oncol. 2012;7:587–94. 11. Proctor MJ, Morrison DS, Talwar D, et al. A comparison of inflammation-based prognostic scores in patients with cancer. A Glasgow Inflammation Outcome Study. Eur J Cancer. 2011;47: 2633–41. 12. Kao SC, Pavlakis N, Harvie R, et al. High blood neutrophil-tolymphocyte ratio is an indicator of poor prognosis in malignant mesothelioma patients undergoing systemic therapy. Clin Cancer Res. 2010;16:5805–13. ¨ zyurtkan MO, Balci AE, Cakmak M. Predictors of mortality 13. O within three months in the patients with malignant pleural effusion. Eur J Intern Med. 2010;21:30–4. 14. Anevlavis S, Kouliatsis G, Sotiriou I, et al. Prognostic factors in patients presenting with pleural effusion revealing malignancy. Respiration. 2014;87:311–6. 15. Al Murri AM, Bartlett JM, Canney PA, Doughty JC, Wilson C, McMillan DC. Evaluation of an inflammation-based prognostic score (GPS) in patients with metastatic breast cancer. Br J Cancer. 2006;94:227–30. 16. Li MX, Liu XM, Zhang XF, et al. Prognostic role of neutrophilto-lymphocyte ratio in colorectal cancer: a systematic review and meta-analysis. Int J Cancer. 2014;134:2403–13. 17. Clive OA, Kahan CB, Hooper CE, et al. Predicting survival in malignant pleural effusion: development and validation of the LENT prognostic score. Thorax. 2014;69:1098–104. 18. Sa´nchez-Mun˜oz A, Pe´rez-Ruiz E, Sa´ez MI, et al. Limited impact of palliative chemotherapy on survival in advanced solid tumors in patients with poor performance status. Clin Transl Oncol. 2011;13:426–9. 19. Dresler CM, Olak J, Herndon JE II, et al. Phase III intergroup study of talc poudrage vs talc slurry sclerosis for malignant pleural effusion. Chest. 2005;127:909–15.
ORIGINAL ARTICLE – THORACIC ONCOLOGY
Prognostic Factors of 30-Day Mortality After Palliative Procedures in Patients with Malignant Pleural Effusion Fernando Conrado Abrao, MD1, Igor Renato Louro Bruno de Abreu, MD2, Mariana Fogarolli, MD2, Giovanna Caxeiro, MD2, Camila Borges Saes Bezerra, MD2, Fernanda Prado de Cerqueira Cesar, MD2, Poline Spitti Rocha, MD2, and Riad Naim Younes, PhD1 Department of Thoracic Surgery, Hospital Sa˜o Jose´, Sa˜o Paulo, SP, Brazil; 2Department of Thoracic Surgery, Hospital Santa Marcelina, Sa˜o Paulo, Brazil
1
ABSTRACT Background. The aim of this study was to identify predictors of 30-day mortality in patients with malignant pleural effusion (MPE) who need pleural palliative procedures. Methods. Data was prospectively collected from our database, between January 2013 and July 2014 on 86 patients with MPE and complete follow-up 30 days after the procedure. The sample was divided into two groups. The first group (G1) included patients who had died up to 30 days after the palliative procedure, and the second group (G2) included patients who survived for more than 30 days after the palliative procedure. The identification of prognostic factors occurred through univariate analysis using Fisher exact test for analysis of categorical variables and the t test for quantitative variables. Subsequently, all variables were used in the multivariate logistic regression analysis. The cutoff values for any significant continuous variables were determined by receiver operating characteristics analysis. Results. There were 24 patients in G1 and 62 patients in G2. Univariate analysis of factors affecting postprocedural survival disclosed nine factors that were associated with significantly reduced postoperative survival. At the multivariate analysis, high levels of white blood cells, (p = .013), low levels of red blood cells (p \ .0001) and protein in pleural fluid (p = .001), and primary lung and gastrointestinal sites (p = .0076) were identified as independent predictors of mortality.
Ó Society of Surgical Oncology 2015 First Received: 30 January 2015 F. C. Abrao, MD e-mail: [email protected]
Conclusions. We identified four factors that are easily recognized in daily practice and can help select patients with low life expectancy. Therefore, invasive procedures and hospitalizations for this subgroup of patients can be prevented.
Malignant pleural effusion (MPE) is common, and given the year-on-year increase in new cancer diagnoses the incidence is set to rise.1,2 MPE represents advanced malignant disease, and current guidelines report median survival rates \1 year and decreased quality of life.3,4 Pleural and oncological treatment options are increasing, and hence more accurate prognosis at presentation may help to identify patients with the poorest prognosis and thus minimize unnecessary procedures and discomfort in their final stages of life.5,6 A few studies have shown predictors of survival in patients with malignant pleural disease, including tumor characteristics, extent of disease, comorbidities, and the composition of the effusion, in addition to inflammationbased scoring systems.7–12 However, to date, no study has attempted to identify predictors of mortality at 30 days, which would greatly help to choose less-invasive palliation options. The aim of this study was to identify predictors of 30-day mortality in patients with malignant pleural effusion who need pleural palliative procedures. MATERIALS AND METHODS A prospectively collected database from our institution, between January 2013 and July 2014, was used to identify patients with MPE submitted to the first pleural palliative procedure. All patients had been followed for a minimum of 1 month or until death.
F. C. Abrao et al.
The ethics committee of the university hospital of Santa Marcelina medical school in Sa˜o Paulo, Brazil, approved this study, under protocol number 39462714.2. Malignant pleural effusion was defined as cases with positive malignant cells in pleural fluid, those with neoplastic pleural infiltration identified in anatomopathological assessment or in patients unfit for pleuroscopy and pleural biopsy, but with metastatic cancer at other sites confirmed by the pathological study, and pleural effusion with no other diagnosed cause after review of the clinical team and pleural fluid analysis. Data on characteristics available preoperatively were collected to allow analysis of prognostic factors, which may influence survival. Collected data included basic demographics, American Society of Anesthesiologists (ASA) physical status classification, performance status according Eastern Cooperative Oncology Group (ECOG) score, hematology parameters, including white blood cells (WBC), number of neutrophils and lymphocytes, neutrophil/lymphocyte ratio (NLR) red blood cells (RBC), and body mass index (BMI) on the day before surgery. The influence of the primary tumor site was also assessed. During the period after the palliative procedure, we evaluated the volume of drained liquid, the type of palliative procedure performed, recurrence of pleural effusion, and the presence of neoplastic cells in pleural fluid, in addition to the biochemical profile including pleural fluid pH, levels of adenosine deaminase (ADA), total protein, albumin, glucose, lactate dehydrogenase (DHL), and the proportion of lymphocytes in pleural fluid (Table 1). Referring physicians and oncologists were contacted to obtain survival data. The palliative method was defined after multidisciplinary discussion. Thoracocentesis was used in those patients who the clinical team thought would have pleural response to chemotherapy, and, therefore, more invasive procedures would not be required. The remaining patients underwent more invasive procedure. In patients considered unfit for general anesthesia, an intercostal drain was inserted under local anesthetic procedure. Once the lung had reexpanded and the drainage reduced, talc slurry was injected into the pleural cavity through the drain. Also in patients considered unfit for general anesthesia, we started using an indwelling pleural catheter in 2014, when it was made available at our institution. Patients treated under general anesthetic procedure underwent an initial bronchoscopy to exclude central airway obstruction. With the patient in the lateral decubitus position, a single port was fashioned to allow inspection of the pleural space with video-assisted thoracoscopy (VAT). After appropriate effusion samples were obtained for cytological examination and culture, multiple large pleural biopsies were taken from any abnormal areas and, if the pleura were uniform, from a number of representative
areas, including the posteroinferior costal pleura. If no preoperative diagnosis was available, these were submitted to frozen-section histopathological examination. If lung reinflation was deemed adequate to allow apposition of the pleural surfaces over most of the hemithorax, then pleurodesis was attempted by insufflation of 4 g of sterile talc. Where the pleural space was obliterated by the tumor, no procedure other than pleural biopsy was performed. In patients undergoing VAT or intercostal drain inserted under local anesthesia, those in whom lung expansion was not greater than half of the hemithorax had the tube removed after 24 h after the procedure and the patient was submitted to monthly revaluations to verify the necessity of new pleural procedure (thoracentesis or indwelling catheter). We divided our sample into two groups. The first group (group I) included patients who had died up to 30 days after the palliative procedure, and the second group (group II) included patients who survived more than 30 days after the palliative procedure. The identification of prognostic factors was performed through univariate analysis, using Fisher exact test for the analysis of categorical variables and t test for quantitative variables. Subsequently, all variables were used in the multivariate logistic regression analysis. This initial model was refined to originate the final model with the significant variables. The p values of .05 or less were considered significant. The cutoff points for any significant continuous variables were determined by receiver operating characteristics (ROC) analysis. Calculations were performed using the statistical software SPSS, version 20.0.
RESULTS A total of 94 consecutive patients underwent the first palliative procedure for MPE during an 18-month period. The most common causes were breast and lung carcinoma. All patient characteristics are shown in Table 1. Eight patients were excluded, as we were unable to get in touch with them 30 days after the pleural procedure, totaling 86 analyzed patients (Table 2). There were 24 patients in group I and 62 patients in group II. The causes of death in group I (perioperative mortality) were pulmonary embolus, in nine; respiratory failure, in 11; hepatic failure, in two; acute abdomen perforation, in two. None of the patients with leukocytosis had pleural empyema at the surgical procedure. The procedures performed were 23 bedside drainages (25.5 %), 20 bedside pleurodesis through a drain (24.5 %), 16 VAT pleurodesis (19.1 %), 17 thoracocenteses (19.2 %), 8 cases of indwelling pleural catheter (9.6 %), and 2 VAT pleural biopsies (2.1 %). Overall median postoperative survival was 127 days.
30-Day mortality After Palliative Procedures TABLE 1 Characteristics of the study population (n = 94) Characteristics
n
%
TABLE 2 Univariate analysis of possible factors related to survival at 30 days (n = 86) Characteristic/group
Gender Female
67
71.3
Male
27
28.7
No (% or SD)
Breast
32
Male
27.7 34.0
17 (27.9)
44 (72.1)
7 (28.0)
18 (72.0)
6 (27.3)
16 (72.7) 26 (83.9)
Primary site
.0463
Gastrointestinal
13
13.8
Lung
Othersa
23
24.2
Breast
5 (16.1)
Gastrointestinal
8 (61.5)
Othersa
5 (25)
ASA 1
1
1.1
2 3
37 49
39.4 52.1
4
7
7.4
Yes (% or SD) 1.0000
Female
26
p value
Gender
Primary site Lung
Survival [30 days
5 (38.5) 15 (75) \.0001
ASA 1
ECOG
0 (.0)
1 (100.0)
2
2 (6.2)
30 (93.8)
3
16 (34.8)
30 (65.2)
4
6 (85.7)
1 (14.3)
0
0 (.0)
4 (100.0)
0
4
4.3
1
27
28.7
2
33
35.1
1
0 (.0)
23 (100.0)
3
22
23.4
2
7 (22.6)
24 (77.4)
4
8
8.5
3
12 (60.0)
8 (40.0)
4
5 (62.5)
3 (37.5) 13 (56.5)
\.0001
ECOG
Procedure Pleurodesis with drainage
23
24.5
VAT pleurodesis
18
19.1
Drainage
10 (43.5)
Drainage
24
25.5
PleureX
4 (50.0)
4 (50.0)
Thoracocentesis
18
19.1
Pleurodesis with drainage
3 (15.0)
17 (85.0)
Vat pleural biopsy
2
2.1
Pleurodesis with vats
2 (12.5)
14 (87.5)
Indwelling catheter
9
9.6
Thoracocentesis
4 (23.5)
13 (76.5)
Vats
1 (50.0)
1 (50.0)
13 (28.3)
33 (71.7)
Pleural histology/cytology Negative Positive
Procedure
50
53.2
Histocytology
44
46.8
Negative
11 (27.5)
29 (72.5)
81.9
Age (years)
Yes
1.0000
59.6 (11.8)
59.6 (14.3)
.7689
18.1
BMI (kg/m2)
22.8 (4.5)
24.7 (4.1)
.2324
Volume (mL)
902.1 (389.7)
970.8 (482.1)
.4738
ADA (IU/L)
9.8 (6.5)
12.9 (10.3)
.2302 .3090
Positive
Recurrence No
.0887
77 17
ASA American Society of Anesthesiologists (ASA) physical status classification, ECOG Eastern Cooperative Oncology Group performance status score
pH
7.6 (.5)
7.7 (.4)
584.7 (646.4)
466.9 (355.0)
.5202
93.0 (79.3)
100.0 (36.3)
.1543
Others: renal, lymphoma, gynecological, kidney, prostate, and thyroid
LDH (IU/L) Albumin (g/dL)
1.9 (.7)
2.8 (.5)
Univariate analysis of factors affecting postprocedural survival verified that primary tumor site and hemoglobin levels, in addition to seven other variables (Table 2), were associated with a significantly reduced postoperative survival. At the multivariate analysis, high levels of WBC (p = .013), low levels of RBC (p \ .0001), and pleural fluid protein (p = .001), as well as lung and gastrointestinal primary site (p = .0076) were identified as independent predictors of mortality (Table 3). ROC analysis identified a WBC value of 9 9 109/L, RBC of 11 g/dL, and total protein of pleural fluid of 3.6 g/dL as the decision-making
Protein (g/dL)
3.3 (1.1)
4.3 (.9)
\.0001
Hemoglobin (g/dL)
10.1 (1.7)
12.0 (1.9)
\.0001
Leukocytes (109/L)
12229 (6100)
8934 (4717)
.0161 .0047
a
Glucose (g/dL)
9
Neutrophils (10 /L) Lymphocytes (109/L) NLR
\.0001
9965 (5770)
6426 (4226)
1172.8 (762.6)
1499.0 (810.3)
.0831
10.2 (7.5)
5.6 (4.8)
.0008
SD standard deviation, ASA American Society of Anesthesiologists (ASA) physical status classification, ECOG Eastern Cooperative Oncology Group performance status score, BMI body mass index, ADA adenosine deaminase, LDH lactate dehydrogenase, NLR neutrophil/lymphocyte ratio a
Others: renal, lymphoma, gynecological, kidney, prostate, and thyroid
F. C. Abrao et al. TABLE 3 Independent predictors of mortality at 30 days at the multivariate analysis Characteristic
Level 2
Odds ratio
Lower 95 %
Upper 95 %
p value
Lung
Breast
17.4239
2.3208
222.1541
Lung
Breast
11.1742
1.2500
167.7473
Lung
Gastrointestinal
1.4158
.1846
11.7608
Gastrointestinal
Breast
12.3070
1.6514
128.1258
Gastrointestinal
Others
7.8926
.9161
98.5765
Breast
Others
1.5593
.1455
15.9397
Protein (g/dL)
.3061
.1226
.6356
.0010
Hemoglobin (g/dL)
.4201
.2311
.6690
\.0001
Leukocytes (109/L)
1.0002
1.0000
1.0003
.0130
A
100
75 50 25 0 0
10
20
75 50 25 0
30
0
10
20
days ≤9000
≤ 9000 : median= – > 9000 : median= 44
30
days ≤11
>9000
Log Rank p= 0.018
>11
Log Rank p= 0.0002 ≤ 11 : median= 30
IC 95% (-; -)
> 11 : median= -
IC95% (0.3 ; 79.7)
IC 95% (11.9 ; 48.1) IC95% (-; -)
B
A
100 100
90
cumulative survival
cumulative survival
FIG. 2 a Kaplan–Meier survival curve comparing patients with total protein from pleural fluid B3.6 g or [3.6 g/dL. b Kaplan–Meier survival curve comparing patients with 0, 1, 2, 3, or 4 prognostic factors determined by multivariate analysis
.0076
B 100
cumulative survival
FIG. 1 a Kaplan–Meier survival curve comparing patients with a white cell count B9 9 109 or [9 9 109/L. b Kaplan–Meier survival curve comparing patients with a red blood cell count B11 or [11 g/dL
cumulative survival
Primary site
Level 1
75 50 25
80 70 60 50 40 30 20
0 0
10
20
days ≤3,6
>3,6
30
10 0 0
10
none
> 3.6 : median= -
two
20
three
30
four
Log Rank p= 0.0001
Log Rank p= 0.0001
≤ 3.6 : median= 28
one
days
IC 95% (17.3 ; 38.7) IC95% (-; -)
none factor :
median= - IC 95% (-; -)
one factor :
median= - IC 95% (-; -)
two factors :
median= 81 IC 95% (0.0 ; 175.8 )
three factors: median= 16 IC 95% (0.0 ; 30.5 ) four factors:
median= 9.5 IC 95% (-; -)
30-Day mortality After Palliative Procedures
threshold, with the highest diagnostic accuracy for predicting death (Figs. 1, 2a). The presence of none, some, or all of factors, which maintain independent significant associations with 30-day mortality after the palliative procedure, stratifies the patient population into five distinct groups. The median survival of patients with all poor prognostic factors is significantly lower than those with none (81 vs. 9.5 days; p \ .001) (Fig. 2b).
DISCUSSION Survival after MPE palliative procedure is important to determine which procedure is more appropriate to avoid exposing patients with low life expectancy to higher morbimortality. Therefore, the more invasive techniques should be reserved for patients with longer survival. The cutoff of mortality at 30 days was chosen because in our country, culturally it is easier to explain and justify less aggressive procedures in patients known to have a life expectancy of 30 days. Additionally, no studies have evaluated mortality in this particular period, to the best of our knowledge. Our study disclosed 4 factors associated with survival \30 days and are easily identifiable: primary pulmonary and gastrointestinal site, serum hemoglobin \11 g/dL, protein concentration in the pleural fluid \3.6 g/dL, and serum leukocyte levels [9 9 109/L. In addition, the combination of these factors also showed a significant impact on patient survival (Fig. 2b). Therefore, it would be reasonable to submit these patients to palliative procedure with multiple pleural aspirations or the use of an indwelling pleural catheter inserted under local anesthesia. In our study, the histology of the primary tumor was an independent prognostic factor, with breast being the histological type of better prognosis and lung, together with gastrointestinal cancer, the histological types of worse prognosis. The only study that evaluated early mortality (3month survival), with a study methodology similar to ours, showed that breast cancer was also associated with longer survival.13 However, studies that evaluated the primary site as an overall survival prognostic factor have shown controversial results. Anevlavis et al. reported ovarian cancer with median survival of 18 versus 15 months for breast cancer (p = .008) being an independent prognostic factor.14 On the other hand, Pilling et al. described malignant pleural mesothelioma and breast cancer as showing a trend toward increased survival when compared with other primary sites, although this was not statistically significant.7 A lower concentration of pleural fluid protein was associated with a lower survival, in our study. Bielsa et al. showed a mean survival of 2.2 months when the pleural fluid total protein value was than 3.85 g/dL, and this was
statistically significant in the multivariate analysis.8 In the ¨ zyurtkan et al., aimed to identify predictors of study by O mortality in patients with MPE at 3 months, the low concentration of protein was only significant in the univariate analysis (p \ .0001) and the cutoff was B3.7 g/dL.13 All these studies included patients who had received previous oncologic treatment; i.e., MPE was a sign of disease progression. On the other hand, Anevlavis et al. studied 90 patients who had received no systemic treatment for cancer and, therefore, had less advanced disease. In this sample, total protein concentration in the pleural fluid was a factor not related to patient survival.14 The explanation may be the advanced stage of cancer, which is strongly associated with the hypoproteinemia and hypoalbuminemia.8,15 Increased WBC and neutrophil levels and an increased N/L ratio were statistically significantly correlated with poor survival in our univariate analysis. Only WBC was an independent factor predicting survival. Conversely to our study, Anevlavis et al. found that N/L ratio, and not WBC, was a predictor of mortality in the multivariate analysis.14 Pilling et al. showed that a leukocyte level [12 9 109/L was an independent predictor of survival.7 These data suggest that systemic inflammation is also an important prognostic factor.12,14,16 This is consistent with growing evidence that systemic inflammation may contribute to both morbidity (in the form of constitutional symptoms) and mortality in cancer and may represent a common pathway in the fatal progression of malignancy.11 Further studies are necessary to understand the mechanisms behind its prognostic value. As for anemia, few studies have associated it with survival in MPE. Pilling et al. showed that hemoglobin \11 g/dL is associated with lower overall survival in patients with MPE, but without statistical significance.7 However, in our study, there was an impact on 30-day survival. These findings seem to support the association between anemia and advanced disease. Performance status has been shown to be associated with prognosis in MPE patients.8,9,14,17 In our study, ECOG was a factor associated with 30-day survival in the univariate analysis, only. Perhaps if our cutoff survival had ¨ zyurtkan et al., we been 3 months as in the study by O would have obtained ECOG as a factor associated with survival, as they reported.13 There is no doubt that survival in patients with high ECOG is lower; however, the median survival of these patients varies, and it is greater than 30 days in the literature, with a 45-day median for ECOG 3 and 33 days for ECOG 3–4.14,18 Perhaps, as an isolated factor, ECOG is not a good parameter for predicting mortality at 30 days. Our study has some limitations. The population we have studied is the one referred to the thoracic surgical unit for consideration of palliation, and it is not disease-specific. We did not evaluate the oncologic treatment of these
F. C. Abrao et al.
patients; thus, we do not know its influence on survival. The group of patients who died at 30 days is much smaller (24 patients) than the group of 62 patients who died after 30 days. However, this disproportion is also found in the literature, in which Dresler et al. reported 17 % of mortality at 30 days.19 We conclude that the primary site, anemia, leukocytosis, and low protein concentration in pleural fluid are associated with 30-day mortality in patients with MPE submitted to the first pleural palliative procedure. The analysis of these factors can identify patients with low survival and, consequently, help physicians choose the most appropriate palliative procedures. DISCLOSURE
The authors declare no conflicts of interest.
REFERENCES 1. The American Thoracic Society. Management of malignant pleural effusions. Am J Respir Crit Care Med. 2000;162:1987–2001. 2. Cancer Statistics Registrations, England (Series MB1): Office of National Statistics, Stationary Office, 2010. 3. Antunes G, Neville E, Duffy J, Ali N; Pleural Diseases Group, Standards of Care Committee, British Thoracic Society. BTS guidelines for the management of malignant pleural effusions. Thorax. 2003;58(suppl II):29–38. 4. Antony VB, Loddenkemper R, Astoul P, et al. Management of malignant pleural effusions. Eur Respir J. 2001;18:402–19. 5. Roberts ME, Neville E, Berrisford RG, Antunes G, Ali NJ; BTS Pleural Disease Guideline Group. Management of a malignant pleural effusion: British Thoracic Society Pleural Disease Guideline 2010. Thorax. 2010;65(Suppl 2):32–40. 6. Tan C, Sedrakyan A, Browne J, Swift S, Treasure T. The evidence on the effectiveness of management for malignant pleural effusion: a systematic review. Eur J Cardiothorac Surg. 2006;29:829–38. 7. Pilling JE, Dusmet ME, Ladas G, Goldstraw P. Prognostic factors for survival after surgical palliation of malignant pleural effusion. J Thorac Oncol. 2010;5:1544–50.
8. Bielsa S, Salud A, Martinez M, et al. Prognostic significance of pleural fluid data in patients with malignant effusion. Eur J Intern Med. 2008;19:334–9. 9. Burrows CM, Mathews WC, Colt HG. Predicting survival in patients with recurrent symptomatic malignant pleural effusions: an assessment of the prognostic values of physiologic, morphologic, and quality of life measures of extent of disease. Chest. 2000;117:73–8. 10. Pinato DJ, Mauri FA, Ramakrishnan R, Wahab L, Lloyd T, Sharma R. Inflammation-based prognostic indices in malignant pleural mesothelioma. J Thorac Oncol. 2012;7:587–94. 11. Proctor MJ, Morrison DS, Talwar D, et al. A comparison of inflammation-based prognostic scores in patients with cancer. A Glasgow Inflammation Outcome Study. Eur J Cancer. 2011;47: 2633–41. 12. Kao SC, Pavlakis N, Harvie R, et al. High blood neutrophil-tolymphocyte ratio is an indicator of poor prognosis in malignant mesothelioma patients undergoing systemic therapy. Clin Cancer Res. 2010;16:5805–13. ¨ zyurtkan MO, Balci AE, Cakmak M. Predictors of mortality 13. O within three months in the patients with malignant pleural effusion. Eur J Intern Med. 2010;21:30–4. 14. Anevlavis S, Kouliatsis G, Sotiriou I, et al. Prognostic factors in patients presenting with pleural effusion revealing malignancy. Respiration. 2014;87:311–6. 15. Al Murri AM, Bartlett JM, Canney PA, Doughty JC, Wilson C, McMillan DC. Evaluation of an inflammation-based prognostic score (GPS) in patients with metastatic breast cancer. Br J Cancer. 2006;94:227–30. 16. Li MX, Liu XM, Zhang XF, et al. Prognostic role of neutrophilto-lymphocyte ratio in colorectal cancer: a systematic review and meta-analysis. Int J Cancer. 2014;134:2403–13. 17. Clive OA, Kahan CB, Hooper CE, et al. Predicting survival in malignant pleural effusion: development and validation of the LENT prognostic score. Thorax. 2014;69:1098–104. 18. Sa´nchez-Mun˜oz A, Pe´rez-Ruiz E, Sa´ez MI, et al. Limited impact of palliative chemotherapy on survival in advanced solid tumors in patients with poor performance status. Clin Transl Oncol. 2011;13:426–9. 19. Dresler CM, Olak J, Herndon JE II, et al. Phase III intergroup study of talc poudrage vs talc slurry sclerosis for malignant pleural effusion. Chest. 2005;127:909–15.
