Targeted detection of amino acid profiles in pleural fluid and serum by liquid chromatography-tandem mass spectrometry for estimating the risks of four aetiologies of pleural effusion: a proof-of-concept study

Abstract

Background

Differentiating pleural effusion remains a challenge in clinical practice. Serum and effusion chemistries help pulmonologists assess the risk of underlying causes and select further diagnostic procedures. Amino acids (AAs) in body fluids are promising tools for estimating the probability of underlying causes. This proof-of-concept study aimed to investigate the accessibility of serum and pleural fluid AAs in distinguishing the four major etiologies of pleural effusions: malignancy, heart failure, tuberculosis, and pneumonia.

Methods

This study prospectively enrolled 153 patients with pleural effusion and unknown causes on admission. The concentrations of 11 AAs in both pleural fluid and serum were measured using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Logistic regression was used to investigate the potential of the AAs panel. The receiver operating characteristic (ROC) curve was used to assess the performance of the AAs panel. Decision curve analysis (DCA) was used to evaluate the net benefit of the panel for patients with pleural effusion.

Results

Our study included pleural effusions caused by heart failure (n = 23, 15%), malignancy (n = 66, 43%), pneumonia (n = 32, 21%), tuberculosis pleurisy (n = 20, 13%), and other causes (n = 8, 5%). Significant differences were observed among various AAs, including alanine and phenylalanine. The areas under the curves (AUCs) for the 11-AA panel in serum to identify four different causes were all ≥0.75. Among them, the AUC for heart failure-related pleural effusion was 0.90 (95% CI: 0.83–0.97). Furthermore, the decision curves for the AAs panels were consistently above the reference lines.

Conclusions

This proof-of-concept study finds that AAs detection in pleural effusion and serum is possible and feasible, but currently has a minor to moderate added diagnostic role.

KEY MESSAGES

• Pleural fluid and serum amino acid levels in patients with pleural effusion are affected by the cause.

• The pleural fluid and serum AA panels could help estimate the risk of the four most common causes of pleural effusion, especially heart failure.

Introduction

Pleural effusion is a clinical sign that can be caused by more than 60 disorders, either in the pleura or systemically [1,2]. The most common causes are malignancy, heart failure (HF), pneumonia, and tuberculosis, and the disease spectrum varies significantly across different regions [1,3]. Determining the etiology is the first step in managing pleural effusion, as it significantly influences treatment selection [2,4]. Despite significant progress in diagnosis over the past few decades, a timely and accurate diagnosis of the underlying cause of a pleural effusion remains a considerable challenge [5]. Simple and inexpensive gold standards, such as cytology and Ziehl-Neelsen (ZN) staining, are typically used as first-line tools for the differential diagnosis of pleural effusions [2]. These two methods have 100% specificity for malignant and tuberculous pleural effusions (MPE, TPE), but their sensitivities are only 58% [6] and 3% [7], respectively. Therefore, complex, expensive, and invasive tools, such as effusion culture, nucleic acid amplification tests (NAATs), imaging-guided pleural biopsy, and thoracoscopy, are needed when first-line tests are negative [8]. Although widely used in clinical practice, these gold standards have limitations. For instance, a positive effusion bacterial culture can diagnose parapneumonic pleural effusion (PPE), but only half of PPE patients are positive [9]. Additionally, effusion culture has a long turnaround time, hindering early diagnosis. Thoracoscopy is the most accurate diagnostic tool for pleural effusions; however, it is an invasive procedure that requires specialized training. Furthermore, 10%–20% of patients remain without a definite diagnosis even when thoracoscopy is employed [10–12]. For these patients, repeated thoracoscopy, cytology, follow-up, assessment of treatment response, or culture is necessary.

Although serum and effusion chemistries are not gold standards for the differential diagnosis of pleural effusions, they are typically used in conjunction with first-line tests to narrow the cause of a pleural effusion. The major clinical value of serum and effusion chemistries is helping pulmonologists estimate the risk of specific types of pleural effusion. Therefore, it has the potential to facilitate the selection of additional diagnostic tests and to shorten the length of hospital stay [13]. The advantages of serum and effusion chemistries include objectivity, minimal invasiveness, short turnaround times, low costs, and easy accessibility [14]. However, the available serum and effusion biochemistries are far from perfect [14], particularly for MPE [15,16] and PPE [17].

Amino acids (AAs) are metabolites that are often altered under disease conditions, and quantifying their serum levels can aid in diagnosing [18]. For example, some studies used serum AAs profiles to diagnose lung adenocarcinoma [19] and active pulmonary tuberculosis [20,21], revealing an accuracy of >90%. Recent studies reported that AAs in pleural fluid possess high accuracy in diagnosing MPE and TPE [22–24]. However, these studies have limitations. The first limitation is the impaired representativeness of the studied cohort. For example, some studies included only MPE and TPE cases [22,25]. Second, most studies focus solely on detecting pleural fluid AAs and do not investigate the diagnostic value of serum or plasma AAs in patients with pleural effusion. Third, most studies have investigated the diagnostic value of AAs for TPE and MPE, whereas reports on their value for estimating the risk of PPE and HF are scarce.

Using liquid chromatography-tandem mass spectrometry (LC-MS/MS), this study aimed to assess the diagnostic accuracy of pleural fluid and serum AAs for estimating the risk of four major types of pleural effusion (MPE, TPE, PPE, and HF). This study complied with the Standards for Reporting of Diagnostic Accuracy Studies (STARD) reporting guideline [26].

Materials and methods

Participants

The participants in this study were from the SIMPLE (A Study Investigating Markers in PLeural Effusion) study, a prospective, pre-registered, double-blind study performed in the Affiliated Hospital of Inner Mongolia Medical University between September 2018 and July 2021, aimed to investigate the diagnostic accuracy of pleural fluid and serum biomarkers [27]. The inclusion and exclusion criteria, reference standards, and design of the SIMPLE study have been previously introduced in our studies [28,29]. The inclusion criterion is patients with pleural effusion and undetermined causes at the time of visiting the emergency or respiratory department. Patients who met any of the following criteria were excluded: with unclear cause even after discharge; pregnant; <18 years; who developed pleural effusion during hospitalization due to medical intervention or trauma; with relapse pleural effusion and the etiologies were clear. This study was approved by the Ethics Committee of the Affiliated Hospital of Inner Mongolia Medical University (No. 2018011). Written informed consent was obtained from all participants. The study adhered to the principles outlined in the Declaration of Helsinki.

Diagnostic criteria and data extraction

MPE was defined by pleural biopsy or effusion cytology. In cases where MPE was highly suspected but cytology was negative, and the patients were unwilling or unable to undergo pleural biopsy, MPE was defined by the evidence of primary cancer and the exclusion of benign pleural effusion. The diagnostic criteria for TPE were pleural biopsy, effusion Mycobacterium tuberculosis (Mtb) culture, and ZN staining. In cases where TPE was highly probable but microbiological evidence was negative, and the patient was unwilling or unable to undergo a pleural biopsy, TPE was diagnosed based on the response to anti-tuberculosis agents with the exclusion of other causes of pleural effusion. The diagnostic criteria for PPE included imaging findings, effusion culture, physical examination, laboratory tests, patient history, and response to antibiotic therapy [17]. HF-related pleural effusion was diagnosed based on clinical symptoms, imaging characteristics, serum NT-proBNP, and responses to diuretics. Two senior physicians (Z.D. Hu and L. Yan) made the final diagnoses, and consensus was used to resolve any disagreements.

We extracted the baseline characteristics from the participants’ medical records. For patients receiving multiple laboratory tests, only the first ones were used for analysis.

Quantifying AAs in pleural fluid and serum

After admission, we collected approximately 10 mL of pleural fluid specimens into an anticoagulant-free tube. The specimens were centrifuged within four hours after collection. The supernatant was aliquoted into EP tubes and stored at −80 °C to −70 °C.

We used the AB SCIEX API3200MD liquid chromatography-tandem mass spectrometry system and the NeoBase^TM Non-derivatized MS/MS Kit (3040-0010Z) to quantify serum and effusion AAs. To determine the analytical imprecision of the assay, we analyzed a mixed pleural fluid specimen 20 times, and the coefficient of variations (CVs) <15% were observed for all AAs. Laboratory technicians who performed the assay were blinded to the participants’ final diagnoses.

Statistical analysis

Continuous data were reported as medians and interquartile ranges (IQR), and normality was assessed using the Kolmogorov-Smirnov test. Categorical data were expressed as absolute numbers and percentages, and the Chi-square or Fisher’s exact test was used for comparisons. Mann–Whitney U-test or independent Student’s t-test was used to compare continuous data between two groups. Logistic regression was used to investigate the performance of the AA panel. Receiver operating characteristic (ROC) curves with areas under the curve (AUCs) and their 95% confidence intervals (95%CIs) were used to evaluate the diagnostic accuracy of AAs [30]. Decision curve analysis (DCA) was used to estimate the net benefit of AA among the participants [30]. A two-tailed p < 0.05 was considered statistically significant.

Results

Baseline characteristics of the participants

A total of 170 participants were included. Of them, 11 participants had no clear diagnosis after discharge, two participants withdrew their consents, and four participants were excluded due to a history of cancer. Therefore, 153 participants were included: 23 with HF, 66 with MPE, 32 with PPE, 20 with TPE, and 12 with other causes. A total of 153 pleural effusion specimens and 96 sera were available. The characteristics of these participants have been reported in our earlier study [31].

Concentrations of AAs and their diagnostic accuracy

Figure 1 shows the levels of 11 AAs in pleural fluid and serum. We categorized the participants into four groups based on their aetiologies. Table 1 summarizes the increased or decreased AAs in each group. Notably, pleural fluid alanine (Ala) levels were significantly increased in MPE and decreased in TPE, making Ala the only AA with the potential to diagnose both diseases. In addition, we failed to observe any AAs that differed between participants with and without PPE.

Figure 1.

Amino acid concentrations in pleural fluid and serum.

A, HF and non-HF; B, MPE and BPE; C, PPE and non-PPE; D, TPE and non-TPE. BPE, benign pleural effusion, PF, pleural fluid.

Table 1.

Differential amino acids between different aetiologies.

Groups	Pleural fluid		Serum
	Increased	Decreased	Increased	Decreased
HF vs non-HF	Met, Cit	—	Gly, Pro, Orn, Met		—
MPE vs BPE	Pro, Ala	—	—		—
PPE vs non-PPE	—	—	—		—
TPE vs non-TPE	—	Ala	Phe		—

Abbreviations: Ala, alanine; BPE, benign pleural effusion; Cit, citrulline; Gly, glycine; HF, heart failure; Met, methionine; MPE, malignant pleural effusion; Orn, ornithine; Phe, phenylalanine; PPE, parapneumonic pleural effusion; Pro, proline; TPE, tuberculous pleural effusion;—represents no amino acid was observed.

Figure 2 lists the ROC curves of the AAs listed in Table 1 in diagnosing HF (2 A), MPE (2B), and TPE (2 C). Table 2 summarizes the AUCs, sensitivities, specificities, positive and negative likelihood ratios (PLRs, NLRs) of the AAs in diagnosing HF, MPE, and TPE. Generally, the diagnostic value of the AAs was fair. The AUCs (95% CIs) of serum ornithine (Orn) and methionine (Met) for diagnosing HF were 0.74 (0.61–0.86) and 0.73 (0.60–0.86), respectively. The AUC of serum phenylalanine (Phe) for diagnosing TPE is 0.77(0.67–0.87).

Figure 2.

ROC curves of investigated amino acids for differentiating pleural effusion. ROC curves diagnosing HF (A), MPE (B), and TPE (C). PF, pleural fluid.

Table 2.

Diagnostic accuracy of investigated amino acids in pleural fluid and serum.

Amino acids	Target disease	AUC (95%CI)	Cut-off (µmol/L)	Sensitivity (95%CI)	Specificity (95%CI)	PLR (95%CI)	NLR (95%CI)
Pleural fluid
Met	HF	0.64 (0.53–0.75)	> 33.29	0.78 (0.58–0.90)	0.52 (0.43–0.60)	1.61 (1.22–2.13)	0.42 (0.19–0.93)
Cit	HF	0.69 (0.58–0.81)	> 42.42	0.61 (0.41–0.78)	0.77 (0.69–0.83)	2.64 (1.68–4.15)	0.51 (0.30–0.85)
Pro	MPE	0.60 (0.51–0.69)	> 209.0	0.61 (0.49–0.72)	0.57 (0.47–0.67)	1.43 (1.04–1.95)	0.69 (0.50–0.94)
Ala	MPE	0.60 (0.51–0.69)	> 370.0	0.61 (0.49–0.72)	0.59 (0.48–0.68)	1.46 (1.07–2.01)	0.67 (0.49–0.92)
Ala	TPE	0.64 (0.51–0.77)	< 258.6	0.40 (0.22–0.61)	0.85 (0.78–0.90)	2.66 (1.36–5.21)	0.71 (0.49–1.01)
Serum
Gly	HF	0.66 (0.54–0.79)	> 362.9	0.94 (0.74–1.00)	0.45 (0.34–0.56)	1.71 (1.36–2.15)	0.12 (0.02–0.86)
Pro	HF	0.68 (0.53–0.83)	> 245.8	0.56 (0.34–0.75)	0.83 (0.74–0.90)	3.33 (1.75–6.36)	0.53 (0.32–0.90)
Orn	HF	0.74 (0.61–0.86)	> 140.9	0.83 (0.61–0.94)	0.59 (0.48–0.69)	2.03 (1.45–2.84)	0.28 (0.10–0.81)
Met	HF	0.73 (0.60–0.86)	> 33.51	0.83 (0.61–0.94)	0.58 (0.47–0.68)	1.97 (1.41–2.74)	0.29 (0.10–0.83)
Phe	TPE	0.77 (0.67–0.87)	> 140.3	1.00 (0.77–1.00)	0.52 (0.41–0.62)	2.08 (1.66–2.59)	0.00 (NA)

Cut-off value set at maximum Youden index.

Abbreviations: Ala, alanine; AUC, area under the curve; CI, confidence interval; Cit, citrulline; Gly, glycine; HF, heart failure; Met, methionine; MPE, malignant pleural effusion; NA, not applicable; NLR, negative likelihood ratio; Orn, ornithine; Phe, phenylalanine; PLR, positive likelihood ratio; PPE, parapneumonic pleural effusion; Pro, proline; TPE, tuberculous pleural effusion.

Diagnostic accuracy of the 11 AA panel

Figure 3 is the ROC curve of the combination of 11 AAs for diagnosing four different types of pleural effusion. The AUCs, sensitivities, specificities, PLR, and NLR of the panels are summarized in Table 3. The diagnostic accuracy of the 11 AA panel was higher than that of a single AA listed in Table 2. It is worth noting that the 11-AA panel in serum had an AUC of 0.90 (95% CI: 0.83–0.97), indicating high performance in estimating the risk of the given causes.

Figure 3.

ROC curves of the AA panel.

ROC curves for the pleural fluid (A) and serum (B) AA panel to differentiate pleural effusion.

Table 3.

Diagnostic accuracy of the AA panel.

Amino acids	Target Disease	AUC (95%CI)	Cut-off value	Sensitivity (95%CI)	Specificity (95%CI)	PLR (95%CI)	NLR (95%CI)
Pleural fluid	HF	0.85 (0.78–0.92)	> 0.14	0.87 (0.68–0.95)	0.76 (0.68–0.83)	3.46 (2.42–4.96)	0.23 (0.09–0.56)
	MPE	0.68 (0.60–0.77)	> 0.52	0.44 (0.33–0.56)	0.85 (0.76–0.91)	2.55 (1.49–4.35)	0.68 (0.54–0.84)
	PPE	0.73 (0.64–0.82)	> 0.17	0.88 (0.72–0.95)	0.56 (0.47–0.65)	1.93 (1.50–2.48)	0.28 (0.12–0.63)
	TPE	0.71 (0.59–0.83)	> 0.16	0.65 (0.43–0.82)	0.73 (0.65–0.80)	2.40 (1.57–3.68)	0.48 (0.26–0.88)
Serum	HF	0.90 (0.83–0.97)	> 0.17	0.94 (0.74–1.00)	0.74 (0.64–0.83)	3.47 (2.30–5.23)	0.15 (0.04–0.56)
	MPE	0.75 (0.65–0.85)	> 0.51	0.55 (0.40–0.70)	0.84 (0.73–0.92)	3.56 (1.83–6.93)	0.53 (0.37–0.76)
	PPE	0.75 (0.63–0.87)	> 0.30	0.61 (0.41–0.78)	0.85 (0.75–0.91)	3.42 (1.89–6.18)	0.48 (0.28–0.80)
	TPE	0.82 (0.72–0.91)	> 0.09	1.00 (0.77–1.00)	0.58 (0.47–0.68)	2.32 (1.71–3.16)	0.13 (0.02–0.85)

Abbreviations: AUC, area under the curve; CI, confidence interval; HF, heart failure; MPE, malignant pleural effusion; NLR, negative likelihood ratio; PLR, positive likelihood ratio; PPE, parapneumonic pleural effusion; TPE, tuberculous pleural effusion.

As shown in Figure 4, the decision curves of the AA panel were all above the reference lines, indicating that the panel can provide net benefits in patients with pleural effusion.

Figure 4.

Decision curves for the AA panel. A, pleural fluid AA panel; B, serum AA panel.

Discussion

This study found that AAs concentrations, whether in serum or pleural fluid, differed among various types of pleural effusions. The pleural fluid and serum AA panels could help estimate the risk of the four most common causes of pleural effusion. Although the accuracy of single AA was limited, their panel showed high accuracy in estimating the risk of specific causes, particularly HF. Our findings indicate that quantifying AA levels represents an alternative tool in estimating the risks of malignancy, HF, pneumonia, and tuberculosis in patients with pleural effusion. The clinical use of AA may facilitate the selection of additional diagnostic procedures, such as bacterial culture or biopsy.

To date, some studies have investigated the pleural fluid [22,23,32–35] and serum [33] AA profiles of patients with pleural effusions. Compared with these studies, our study has its advantages. First, most previous studies used ¹H nuclear magnetic resonance (¹H NMR) to determine AA [23,32,33,35], while we used the LC-MS/MS with a commercial kit. The technical advantages of the assay include high functional sensitivity and standardization, and low cost [36,37]. These advantages facilitate the generalizability of our findings in clinical practice. Second, the etiologies of pleural effusion encompass more than 60 disorders, with malignancy (MPE), HF, pneumonia (PPE), and tuberculosis (TPE) accounting for approximately 80% of the cases. The disease spectrum in previous studies only encompassed two or three types of pleural effusion (e.g. TPE and MPE [22], MPE, TPE, and PPE [32,33]). This two-gate design may impair the representativeness of participants, bias findings, and, consequently, the generalizability of the findings [38]. By contrast, our prospective study is a one-gate design. The participants in our study are more representative, and the findings are thus more generalizable. Differences in participants and AA assays between our study and previous studies may partly explain the discrepancies in altered AAs. Third, the previous study focused only on identifying differentiated AAs among patients with pleural effusion [23,34,39] or on evaluating the diagnostic accuracy of a single AA [32,33]. By contrast, our study established two panels of effusion or serum AAs and showed that these panels can improve the ability of AAs to estimate the probability of specific types of pleural effusion. Taken together, this work extends knowledge of AA’s ability to differentiate pleural effusion.

The mechanisms underlying the altered AA profile in effusion and serum remain unknown. It may be released by pleural mesothelial cells, immune cells in the effusion, or locally by cancer cells. In addition, it can also be passively diffused from the serum to the pleural space. The disorder in other systems or organs may also change the serum AA profile, and consequently affect the AA profile of pleural fluid. Our study revealed some altered serum or effusion AAs in different types of pleural effusion. These findings may provide a novel insight into the pathogenesis of MPE, PPE, HF, and TPE. Further studies are needed to address the sources and molecular mechanisms underlying the altered AA profile.

Although single AA was inadequate for differentiating pleural effusion, the AA panel showed high accuracy. These findings suggest that multiple AA detection represents a useful tool for optimizing the diagnostic pathway of pleural effusion. Notably, the panel does not increase costs because LC-MS/MS is a high-throughput assay, and all 11 AAs can be determined in a single specimen. If the algorithms under the diagnostic model were incorporated into the laboratory information system, the risk of a given type of pleural effusion could be easily calculated, facilitating pulmonologists’ further diagnostic planning. With this cohort, our previous study indicated that adenosine deaminase (ADA) helped diagnose TPE, with an AUC of 0.80 [39]. In addition, the serum NT-proBNP helped diagnose HF, with an AUC of 0.83 [40]. This study found that the AUCs of the serum and effusion AA panels were 0.90 and 0.85 for HF, and 0.82 and 0.71 for TPE. The AUCs of AA panels were comparable to those of ADA and NT-proBNP, with no statistical significance (data not shown). Therefore, we believe the AA panel’s performance is comparable to that of ADA and NT-proBNP, two guideline-endorsed biomarkers.

It should be emphasized that the role of effusion or serum chemistries is to help pulmonologists estimate the risk of MPE, thereby potentially improving diagnostic efficiency. They cannot replace cytology or histological confirmation, especially in the era of molecular and immunotherapy. This study revealed that the performance of the serum and effusion AA panels for MPE is modest, with AUCs of 0.75 and 0.68, respectively. These results indicate that the AA panel’s performance in estimating MPE risk is limited and may not alter the current clinical pathways for diagnosing MPE. However, a cutoff value of 0.09 for the serum AA panel yielded a sensitivity of 1.00 and an NLR of 0, suggesting that it may serve as a rule-out test for TPE and modify the current clinical diagnostic pathway. For patients with values < 0.09, Mycobacterium tuberculosis culture or NAAT might be avoided.

The AUCs of the AA panels for identifying MPE, PPE, and TPE were < 0.82, indicating limited value. However, the AUCs of the serum and effusion AA panels were >0.85, suggesting they have high value in estimating HF risk. Currently, the most powerful diagnostic tools for HF-related pleural effusion are serum or effusion NT-proBNP [13,41], Porcel’s score[42], and BANCA (Bilateral effusions on chest radiographs, Age, NT-proBNP levels in pleural fluid, Cholesterol in pleural fluid, and Albumin gradient) score [43]. Compared with these tools, the primary advantage of the AA panel is its low cost. In China, the cost of the AA panel is approximately 7$, while the cost of NT-proBNP is 17$. In addition, the AA panel calculation is simple because its algorithm can be incorporated into a local laboratory information system. Because our participants had some missing data on effusion chemistries, particularly albumin gradient, we were unable to compare the performance of the AA panels with these tools. Further studies are needed to address this issue.

Our study has some limitations. The first limitation is the monocenter design conducted in an area with a high tuberculosis prevalence (13%). Extending our findings to areas with a low tuberculosis prevalence should be done cautiously. Second, given the single-center design and small sample size, variable selection was not performed in the logistic regression model, and external validation was not possible. Model performance is mainly determined by two metrics, named bias and variance. Bias represents diagnostic accuracy, while variance reflects the model’s generalizability to other populations. A trade-off exists between these two metrics. The use of more variables may reduce bias but increase variance, and vice versa. As a result, the diagnostic performance of the AA panel in this study might have been overestimated. Future studies should use appropriate variable selection methods, such as the Least Absolute Shrinkage and Selection Operator (LASSO) regression or the Boruta algorithm, to build more robust models and balance bias and variance. The small sample size also hinders our ability to investigate the added value of AAs to promising biomarkers endorsed by the guidelines [44,45], such as ADA and NT-proBNP. The small sample size also prevented us from performing subgroup analyses to investigate the value of AAs in specific patient groups, such as cytology-negative patients. Third, the effusion and serum specimens used in our study have been stored at −80 and −70 degrees for approximately four years. A previous study indicated that some, but not all, of the AAs investigated in our study were unstable under such conditions for 5 years [46]. However, the concentration changes, expressed as percentages, are all <25% and are statistically significant [46]. Considering the imprecision of LC-MS/MS is around 10% and that individuals also exhibit biological variation[47], we believe that the concentration changes at −80 degrees have a minor effect on the reliability of the findings. Fourth, some patients with MPE and TPE were diagnosed clinically in the absence of pathological, cytological, or microbiological evidence. The decision to exclude or include these cases is a double-edged sword. Exclusion would improve diagnostic reliability but at the cost of reduced generalizability, whereas inclusion would enhance population representativeness but decrease reliability. After weighing these trade-offs, we decided to include these patients. Diagnoses were established by two senior physicians based on comprehensive clinical assessment, with particular emphasis on treatment response, to preserve population representativeness while minimizing reliability loss. In light of these limitations, the present study should be considered a proof-of-concept study, and the proposed concept warrants further investigation in future studies.

Conclusions

In summary, this proof-of-concept study finds that AAs detection by LC-MS/MS has a minor to moderate diagnostic value for MPE, PPE, and TPE. However, simultaneously determining AAs in serum and effusion, along with the AA panel, represents a potential tool for improving diagnostic pathways for HF-related pleural effusion. Due to the proof-of-concept nature of this study, our findings need to be validated in future studies with larger sample sizes.

Funding Statement

This work was supported by the Inner Mongolia Medical University Affiliated Hospital Talent Training Project - Sailing Series (HH202401).

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.