Pleural fluid biomarkers to predict response to pleurodesis or indwelling pleural catheter for malignant pleural effusion: still a long way to go

Abstract

This letter summarizes the current literature on pleural fluid biomarkers that predict treatment response to chemical pleurodesis and indwelling pleural catheters in patients with malignant pleural effusion. We found evidence on pleural fluid biomarkers to be limited to a few small studies, where the biomarkers prognostic value was not tested in separate validation cohorts. Although low pleural fluid LDH, high glucose, and high pH levels are associated with pleurodesis success, no clinically valuable biomarker has been identified. Hence, further research on the inflammatory processes of malignant pleural effusion and during definitive pleural procedures is needed, as well as proper testing of potential biomarkers in separate model derivation and validation cohorts.

Introduction

Malignant pleural effusion (MPE) denotes the presence of fluid in the pleural space caused by cancer [1] and is associated with a high symptom burden, reduced quality of life, and shortened life expectancy [2]. Therefore, early and effective treatment is required. The principles of MPE treatment are two-fold: reduction of the tumor burden by systemic anticancer treatment; and reduction of MPE symptoms by pleural interventions such as repeated thoracentesis, talc pleurodesis, indwelling pleural catheter (IPC), or a combination of IPC and pleurodesis [3]. Talc pleurodesis and IPC are comparable in terms of symptom relief. However, talc pleurodesis requires in-hospital admission for 2–3 days and has a success rate of approximately 70% for complete or partial reduction of MPE [4, 5]. An IPC is inserted at an outpatient clinic, and the drain is emptied regularly at the patient’s home, thereby maximizing the time outside the hospital environment. IPC is preferred in cases of large pleural fluid production, the presence of a trapped lung, or failed pleurodesis [6]. Spontaneous pleurodesis occurs in around 24% of patients treated with IPC [7, 8] compared to 43% of patients treated with talc pleurodesis via IPC [7]. Owing to the modest success rates of pleurodesis and IPC and the low life expectancy of patients with MPE, selecting the optimal treatment for each patient is essential. Therefore, we examined the current literature on pleural fluid biomarkers that predict treatment response to chemical pleurodesis and IPC in patients with MPE.

A literature search was conducted in October 2024 (updated in May 2025) by searching Pubmed.gov with the following search strings: ”pleurodesis AND prediction AND pleural fluid” and ”indwelling pleural catheters AND prediction”. This study evaluated biomarker model development, with particular emphasis on the inclusion of established clinical predictors (e.g., trapped lung and daily pleural fluid production volume), incorporation of a validation cohort distinct from the model cohort, and examination of previously published models.

Biomarkers for chemical pleurodesis outcome

Most studies on pleural fluid predictors have focused on the success rate of chemical pleurodesis. In 2020, Hassan et al. [9] published a comprehensive systematic review and meta-analysis of the predictors of chemical pleurodesis outcomes in patients with MPE. This systematic review encompassed 34 studies, and the meta-analysis incorporated data from 26 studies comprising 3,264 participants. Pleural fluid lactate dehydrogenases (LDH), glucose, pH, and neutrophil levels were identified as prognostic biomarkers. We further identified studies that focused on specific proteins. Table 1 provides an overview of the findings, including the assessment of biomarker model quality.

Table 1.

Pleural fluid biomarker for pleurodesis success: biomarkers identified, number of studies, positive association to successful intervention, and status on biomarker model validation

	N studies	Association to successful intervention	Meta analysis	Adjusted for known clinical predictors	Model validated
Biomarker for Pleurodesis success
LDH: low	3	Positive	-728 IU/L (95% CI -1499 to -43)	No	No
Glucose: high	5	Positive	36.6 mg/dL (95% CI 16.3–56.2)	No	No
pH: high	6	Positive	0.12 (95% CI 0.11–0.14)	No	No
Neutrophil count	2	Contradictory	-	No	No
CRP	3	Contradictory	-	No	No
Protein	2	Contradictory	-	No	No
ADA: high	2	Positive	-	No	No
Osteopontin	1	None	-	No	No
uPA	1	None	-	No	No
VEGF	2	None	-	No	No
sVEGFR	1	None	-	No	No
D-dimer	1	None	-	No	No
sCD163: high	1	Negative	-	No	No
Survivin: high	1	Negative	-	No	No
Biomarker for IPC success
Protein: high	2	Contradictory	-	No	No
Albumin: high	1	Positive	-	No	No
Eosinophils: high	1	Positive	-	No	No
CRP	1	None	-	No	No
LDH	1	None	-	No	No
VEGF: high	1	Positive	-	No	No
β-FGF: high	1	Postive	-	No	No
IL-8	1	None	-	No	No
TGF-β	1	None	-	No	No

ADA: adenosine deaminase, CD163: cluster of differentiation 163, CRP: C-reactive protein, FGF: fibroblast growth factor, IL-8: interleukin-8, LDH: lactate dehydrogenases, sCD163: soluble CD163 scavenger receptor, sVEGFR: soluble vascular endothelial growth factor receptor, TGF-β1: transforming growth factor-β1, uPA: urokinase-type plasminogen activator, VEGF: vascular endothelial growth factor

Pleural fluid LDH has been investigated as a biomarker in one retrospective [10] and two prospective studies [11, 12]. The pooled results from the 232 patients showed that lower LDH levels were associated with pleurodesis success with a mean difference between the success and failure groups of -728 IU/L (95% CI -1499 to -43, p = 0.064, I² 95.5%, p < 0.001)⁹. Neither study included clinical variables or a validation cohort (Table 1).

Pleural fluid glucose levels as a biomarker is addressed by Hassan et al. [9], who analyzed data from five studies and 709 patients (two retrospective [10, 13, two prospective11, 12], and a meta-analysis by Heffner et al. [14]). They found that higher glucose levels favored pleurodesis success with a mean glucose difference of 36.6 mg/dL (2 mmol/L, 95% CI 16.3–56.2 mg/dL, p < 0.001, I2 93%, p < 0.001). However, there was a lack of agreement between studies regarding the cutoff values. Two of the included studies suggested a cutoff value for pleural fluid glucose to predict pleurodesis failure. A meta-analysis by Heffner et al. (n = 433)¹⁴ suggested a glucose threshold of < 72 mg/dL (4 mmol/L) with an area under the curve (AUC) of 0.66 (95% CI 0.61–0.72), indicating limited diagnostic efficacy. Lastly, Pantazouplos et al. [10] reported an optimal cutoff of glucose < 65 mg/dL (3.6 mmol/L) with an AUC of 0.81 and a sensitivity and specificity of 90.7% and 76.8%, respectively (p < 0.001); however, this retrospective study only included 162 patients. Neither study included clinical variables or a validation cohort (Table 1).

High pleural fluid pH [9], examined in six studies [10, 11, 12, 13, 14, 15], was associated with pleurodesis success, yet with a modest mean difference in pleural fluid pH between the pleurodesis success and failure groups of 0.124 (95% CI 0.107–0.140, p = 0.008, I² 0%, p = 0.473). Neither study included clinical variables or a validation cohort (Table 1).

We found only two studies that assessed the predictive ability of the pleural fluid neutrophil count for pleurodesis success [16, 17]. A prospective study by Psathakis et al. [17], which included 168 patients, reported that pleural fluid neutrophils increased more intensely after talc poudrage in patients with successful pleurodesis than in those with failed pleurodesis (p = 0.005); however, there were no significant differences in baseline levels. Moreover, a prospective study by Habal et al. [16] including 114 patients, found a significantly higher proportion of pleural fluid granulocytes in the group with successful talc poudrage than in the failed group. They found no difference in the baseline pleural fluid C-reactive protein (CRP) levels. In contrast, Lapidot et al. [18] found increased baseline pleural fluid CRP levels in patients with successful pleurodesis, in a small prospective study involving 34 patients. They hypothesized that increased CRP levels represent existing inflammation, which may amplify the local inflammation induced by pleurodesis. However, a small prospective study of 40 patients [12] reported the opposite, as a significantly lower pleural fluid CRP was associated with successful pleurodesis (2.62 mg/L ± 0.57 vs. 4.73 mg/L ± 1.63, p < 0.05). In contrast to Lapidot et al. they hypothesized that high CRP levels indicate neoplastic progression, larger pleural fluid production, and thus a decreased effect of talc pleurodesis. Therefore, the role of pleural fluid neutrophils and CRP as predictive biomarkers of pleurodesis outcomes remains unclear. Neither study included clinical variables or a validation cohort (Table 1).

There is a paucity of literature concerning the association between pleural fluid protein levels and pleurodesis outcomes. The retrospective study by Pantazopoulos et al. [10] found the level of pleural fluid protein to be significantly lower in patients with pleurodesis failure (34.2%) (median [IQR] 3.6 g/dL (3.2–4.4) vs. 4.5 g/dL (4.1–4.9, p < 0.001). They reported that pleural fluid protein < 4.2 g/dl had an AUC of 0.73 to predict pleurodesis failure. In contrast, Lapidot et al. [18] did not identify pleural fluid protein as a predictor of pleurodesis outcomes. Neither study included clinical variables nor a validation cohort, see Table 1.

Other pleural fluid biomarkers for pleurodesis outcomes have been suggested in individual studies. A prospective study by Yildirim et al. [19], which included 60 patients, found that baseline pleural fluid adenosine deaminase (ADA) levels in patients with successful pleurodesis were significantly higher than those in patients with unsuccessful pleurodesis (p < 0.001). This result was interpreted as a sign of increased lymphocyte activity in patients with successful pleurodesis. The optimal cutoff was 17.5 U/L (AUC = 0.873; sensitivity = 77.6%; specificity = 90.9%). Subsequently, the same research group [20] published a larger prospective study that included 84 patients from the same institution and again identified ADA level as an independent predictor of talc pleurodesis outcomes. However, it is not completely clear whether the study populations were different, as the recruitment periods and inclusion criteria of the two studies overlapped [19, 20]. Thus, pleural fluid ADA may be useful as an independent predictor; however, larger studies, including subgroup analyses of various cancer types, are needed. Neither study included clinical variables or a validation cohort (Table 1), which is also true for the remaining studies presented below.

In a prospective study conducted in 2008, Hsu et al. [21] investigated the value of osteopontin, vascular endothelial growth factor (VEGF), and urokinase-type plasminogen activator (uPA) as predictors of pleurodesis outcome in 61 patients but did not identify any predictive value for these biomarkers. VEGF and its soluble receptor sVEGFR-1 were also investigated in 2012 by Hooper et al. [22], who found no association between baseline pleural fluid VEGF, sVEGFR-1, and pleurodesis outcomes. Thus, there is no evidence to suggest that VEGF is a clinically useful predictor.

Psathakis et al. [17] found that pleural fluid D-dimer levels decreased significantly 24 h after pleurodesis in the group with successful pleurodesis, reflecting intrapleural coagulation. However, the baseline D-dimer level was not associated with pleurodesis outcomes and was not useful as a predictor.

In a retrospective study including 114 patients, Habal et al. [23] found significantly higher levels of soluble CD163 scavenger receptor in pleural fluid before talc pleurodesis in patients with pleurodesis failure (P0 = 0.00024). However, no prospective validation was performed.

Pleural fluid survivin, an inhibitor of apoptosis related to increased tumor aggressiveness, was investigated by Arellano-Orden et al. [24] as a predictor of pleurodesis success. In 84 patients who underwent talc pleurodesis, higher baseline survivin levels were associated with a significantly lower pleurodesis rate (p = 0.004). However, the optimal cutoff of 30 pg/mL predicted pleurodesis failure with 54% sensitivity and 79% specificity, thus being a predictor of limited value.

Biomarkers for indwelling pleural catheter outcome

Few studies have investigated pleural fluid biomarkers for treatment response to IPCs, and no study has included both model and validation cohorts. A large retrospective study from 2018 by Li et al. [25], including 1071 IPCs in 956 patients, found an association between increased rates of auto-pleurodesis and increased pleural fluid total protein level (OR 1.03, 95% CI 1.01–1.06), albumin level (OR 1.07, 95% CI 1.03–1.12) and percent eosinophils (OR 1.04, 95% CI 1.00-1.07) [25]. However, the optimal threshold values had only modest predictive value (protein ≥ 44 g/L, AUC 0.69 [95% CI 0.65–0.72]; albumin ≥ 25 g/L AUC 0.71 [95% CI 0.68–0.75]; % eosinophils ≥ 6% AUC 0.55 [95% CI 0.52–0.58]). Another retrospective study by Porcel et al. [26] in 2020, including 336 IPCs, found no association between pleural fluid protein, CRP, or LDH levels and auto-pleurodesis. In a small, exploratory study other pleural fluid inflammatory biomarkers have been suggested as potential predictors of pleurodesis in patients with IPC, including interleukin-8 (IL-8), VEGF, β-fibroblast growth factor (β-FGF) and transforming growth factor-β (TGF-β) [27, 28]. These assays are not routinely available in most facilities, and neither study included clinical variables or a validation cohort (Table 1).

The Porcel et al. study also addressed biomarkers for complications to IPC and found that pleural fluid CRP < 15 mg/L before IPC insertion was an independent predictor of IPC-related pleural infection (OR = 4.42 95%CI, 1.26–15.52) [26].

Discussion

This research letter highlights the absence of suitable pleural fluid biomarkers to support clinical decision-making in the management of MPE. This is particularly relevant to treatment with IPCs, presumably due to the long-standing use of talc slurry pleurodesis via chest tube or thoracoscopic talc poudrage as first-line interventions. The current evidence on biomarkers is limited to a few small prospective studies with substantial heterogeneity in the definition of pleurodesis success (radiographic vs. patient symptoms) and the type of pleurodesis agent. Overall, the available studies do not adhere to gold standards for biomarker development, such as presentation of prognostic performance in separate model derivation and validation cohorts, and presentation of the added prognostic value to already known clinical biomarkers such as trapped lung or daily MPE fluid volume production [29].

Although there is evidence that low pleural fluid LDH, high glucose, and pH levels are associated with pleurodesis success in several small studies, none of them have shown promise as independent predictive biomarkers. High pleural fluid LDH, low glucose, low pH, and high protein levels reflect increased metabolic activity and thus high cancer cell activity, which may hinder the pleurodesis. However, they may also reflect more general inflammation, which could induce local inflammation during pleurodesis. Hence, there is a need for further understanding of the inflammatory processes involved in MPE and the response to definitive pleural procedures in these patients. The identification of key mechanisms for success, failure, or adverse events will improve patient-tailored MPE management, and could potentially lead to the development of targeted drug therapies to control MPE production [30]. Currently, clinicians must still rely on clinical data derived from chest radiographs and measurement of fluid volume output to provide imperfect clues to the predicted outcome of pleurodesis.

Abbreviations

ADA

Adenosine deaminase

AUC

Area under the curve

CRP

C-reactive protein

IL-8

Interleukin-8

IPC

Indwelling pleural catheter

LDH

Lactate dehydrogenases

MPE

Malignant pleural effusion

sVEGFR-1

Soluble vascular endothelial growth factor receptor 1

TGF-β

Transforming growth factor-β

VEGF

Vascular endothelial growth factor

uPA

Urokinase-type plasminogen activator

β-FGF

β-fibroblast growth factor

Author contributions

K.F. performed the litterature search. K.F., S.A. and U.B. wrote the first draft for the main manuscript text. K.F. and U.B. prepared Table 1. J.P., F.M and J.S. contributed with substantial rivisions of the manuscript. All authors reviewed the manuscript.

Funding

The authors received no funding for the development of this manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.