Diagnostic utility of pleural fluid adenosine deaminase in tuberculous pleural effusion in patients with exudative effusion: a retrospective multicenter study (ADATPE).

Worldwide, tuberculosis (TB) remains a significant global health challenge, with an estimated 10 million cases reported in 2020 according to the World Health Organization (WHO). The global incidence rate is approximately 134 cases per 100,000 population, with higher rates in low- and middle-income countries. In Malaysia, the TB incidence rate is also concerning, with around 122 cases per 100,000 population as of 2023. The government reported 26,781 TB cases in 2023, reflecting a 5.5% increase from the previous year. Despite ongoing efforts to control the disease, TB continues to be a public health threat in Malaysia ( 1 ). Tuberculous pleural effusion (TPE), a common complication of TB, occurs when the infection spreads to the pleural space, accumulating fluid in the lungs ( 2 ). It is estimated that up to 20% of patients with pulmonary TB may develop pleural effusion. The incidence of pleural effusion is higher in patients with untreated or inadequately treated TB. Pleural effusion may occur during active tuberculosis infection or may develop later as a consequence of previously untreated or inadequately treated tuberculosis, leading to complications such as chronic respiratory symptoms, lung damage, or fibrosis ( 3 , 4 ). Early diagnosis and treatment of TB are crucial in preventing the progression to pleural effusion and associated complications ( 3 ). Although TPE is a common clinical entity, it can be challenging to diagnose. The most accurate, but also the costliest, method for diagnosing TPE is thoracoscopic pleural biopsy. It demonstrates a sensitivity of up to 100% in diagnosing TPE, providing a more reliable method for identifying the disease ( 5 ). However, most high-TB-incidence regions lack the infrastructure required for routine thoracoscopy and have limited financial resources, highlighting the need for alternative, cost-effective diagnostic methods for pleural effusions ( 6 , 7 ). Pleural fluid for conventional smear microscopy to diagnose TPE had low yield rates of under 10% ( 8 ). Pleural fluid cultures using solid media are positive in less than 40% of cases ( 3 ). The Xpert Mycobacterium tuberculosis /rifampicin resistance (MTB/RIF) assay, an automated molecular diagnostic test, has significantly enhanced the detection of TB and rifampicin resistance, with a sensitivity of 84.7% and a specificity of 98.4%. However, its sensitivity remains suboptimal in individuals with paucibacillary disease or human immunodeficiency virus (HIV) co-infection, limiting its utility in cases of extrapulmonary or sputum smear-negative TB. This is particularly critical for HIV-positive individuals and children, as TB is often challenging to diagnose in these populations and is associated with a high morbidity rate ( 9 , 10 ). In contrast, the newer GeneXpert Ultra test demonstrates improved sensitivity (90.9%) with specificity of 95.6%, offering better detection capabilities, particularly in low-bacterial-load samples and HIV-infected patients ( 11 , 12 ). In countries with moderate to high TB incidence, pleural fluid adenosine deaminase (pfADA) measurement is commonly used to investigate undiagnosed pleural effusions or support pleural fluid analysis when TPE is suspected. Adenosine deaminase (ADA) is a purine catabolic enzyme that catalyzes the conversion of adenosine to inosine and is particularly abundant in lymphoid tissue. Many studies have suggested that pleural fluid pfADA is helpful in the diagnosis of TPE ( 7 , 13 - 15 ). A recent study indicates that combining pleural fluid ADA with lactate dehydrogenase (LDH)-based ratios or interferon-γ assays improves diagnostic discrimination between tuberculous and non-TPEs and enhances specificity in patients with elevated ADA levels ( 16 ). The merits of pfADA include its low cost, short turnaround time, and high sensitivity (93%) and specificity (90%) ( 7 , 17 ). In high TB prevalence settings, patients with lymphocyte-dominant exudative pleural effusion could be diagnosed as TPE when pfADA is over 40 IU/L based on the excellent diagnostic accuracy of ADA in these settings ( 4 , 18 - 20 ). However, the role of pfADA should be re-established when its epidemiologic status changes, and to provide local data regarding the sensitivity, specificity, and cutoff level of pfADA in diagnosing TPE. Studies in Hong Kong and East Malaysia have shown that the diagnostic accuracy of pleural fluid ADA can be further optimised by establishing a local pfADA cutoff value ( 21 , 22 ). In addition, the pleural fluid LDH level is elevated in approximately 75% of cases of TPE, with levels commonly exceeding 500 IU/L ( 20 , 23 ). Wang et al. found that pleural fluid LDH is a valuable biomarker for differentiating TPE from parapneumonic effusion (PPE), with LDH levels being significantly lower in TPE patients. They also observed a greater difference when comparing TPE with complicated parapneumonic effusion (CPPE) and empyema. Due to the limitations of using pleural fluid ADA and LDH levels alone, they combined both parameters to develop a more reliable predictor of TPE, finding a significantly lower LDH/ADA ratio in the TPE group compared to the PPE group ( 23 , 24 ). Furthermore, there is no standardised value of the pleural fluid LDH/ADA ratio in the diagnosis of TPE. This study aimed to optimize the utility and validity of pfADA in diagnosing TPE by: (I) determining the local cutoff level of pfADA for diagnosing TPE; (II) assessing the sensitivity and specificity of the current pfADA cutoff value of 40 U/L in diagnosing TPE; and (III) investigating the association between pleural fluid LDH and the ADA ratio in the diagnosis of TPE. We present this article in accordance with the STARD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0015/rc ).

The ADATPE study (The Diagnostic Utility and Validity of Pleural Fluid Adenosine Deaminase Levels in the Diagnosis of Tuberculous Pleural Effusion in Patients with Exudative Effusion) is a retrospective, observational, multicenter study conducted across three centers in Malaysia: Hospital Canselor Tuanku Muhriz (HCTM) of National University of Malaysia [Universiti Kebangsaan Malaysia (UKM)], University Malaya Medical Centre (UMMC), and Hospital Sultanah Bahiyah, Alor Star. This study was approved by the National Medical Research Register (NMRR ID-24-01591-ZKB) on the 22 nd July 2024, and the Research Ethics Committee of Universiti Kebangsaan Malaysia (FF-2023-131). All participating institutions were also informed of and agreed to the study. This study is in accordance with the Helsinki Declaration and its subsequent amendments. Patient consent was not required for this study, as it is a retrospective analysis, and the Research Ethics Committee at Universiti Kebangsaan Malaysia waived the need for consent. Data were collected using a universal random sampling method over a six-year period, from January 2019 to April 2024. All patients aged >16 fulfilling Light’s criteria for exudative effusion (pleural fluid protein/serum protein >0.5; pleural fluid LDH/serum LDH >0.6; pleural fluid LDH > two-thirds of the standard upper limit of serum LDH) were included in the study. The study excluded patients with incomplete pfADA or other laboratory results, as well as those whose medical records could not be accessed due to missing files, technical issues, or other barriers to data retrieval. Patients with transudative pleural effusions, as defined by Light’s criteria, were excluded from this study, which focused on exudative effusions, particularly those with TPE. The sample size calculation was based on a sensitivity of 0.88 and specificity of 0.92 for pfADA in diagnosing TPE, as reported in a previous study by Huan et al. ( 22 ). To achieve a precision of 0.075 for both sensitivity and specificity, a total sample size of at least 166 was required; however, we obtained 276 samples. The following data were collected from study subjects from Bed Head Ticket (BHT) and Electronic Medical Record (EMR): demographic data, including age (at the time of pfADA testing), gender, race, smoking status, and associated comorbidities [diabetes mellitus, hypertension, dyslipidaemia, cardiovascular disease, chronic kidney disease (CKD), chronic lung disease, chronic liver disease, and haematological malignancy]. Laboratory results included white cell counts, serum protein, albumin, and LDH. The nature of pleural fluid (exudative or transudative based on Light’s criteria), pleural fluid or biopsy bacteriology findings (acid-fast stain, TB culture and sensitivity, or Xpert MTB/RIF), pleural fluid cytology results, pleural fluid or pleural biopsy histopathology, pleural fluid ADA levels, and definitive diagnosis were also recorded. Pleural fluid ADA levels were measured using spectrophotometry. All tests from HCTM and UMMC were sent to Lablink Medical Laboratory in Kuala Lumpur, Malaysia, and subsequently analyzed at Dr. Lal Path Labs in India. Pleural fluid ADA samples from Hospital Sultanah Bahiyah, Alor Setar, were sent to the National Public Health Laboratory, Ipoh, Perak, which were analysed via colorimetric assay. Patients were categorized into three groups based on a composite reference standard comprising clinical, laboratory, histopathologic, and radiologic examinations: ❖ Group 1: definite TPE. Patients with at least one positive MTB culture and/or acid-fast stain and/or Xpert MTB from pleural fluid or biopsy specimens and/or caseating granulomatous inflammation suggestive of TB on histological examination of pleural biopsy tissue ( 3 , 8 ). ❖ Group 2: clinical TPE. Patients who did not fulfill the criteria for bacteriologic and histopathologic confirmation but were diagnosed with TPE by a respiratory physician based on clinical symptoms, radiologic imaging, thoracoscopy findings suggestive of TPE, epidemiological risk factors, and subsequent clinical response to empirical anti-TB therapy when other diagnoses had been reasonably excluded. No single parameter, including pleural fluid ADA level, was used in isolation to establish the diagnosis or to initiate treatment ( 25 , 26 ). ❖ Group 3: non-TPE. Patients with pleural effusion with no microbiological or histological evidence of Mycobacterium tuberculosis and/or for whom an alternative diagnosis was available. Data were analysed using the SPSS software. The receiver operating characteristic (ROC) curve and the Youden Index were used to evaluate the accuracy of pfADA and to determine the optimal local cut-off value. Continuous variables were presented as mean ± standard deviation for normally distributed data and median [interquartile range (IQR)] for non-normally distributed data. Normally distributed continuous variables were compared using the independent t-test, while non-normally distributed variables were compared using the Mann-Whitney U test. Categorical variables were analysed using the chi-squared test or Fisher’s exact test. A P value of less than 0.05 was considered statistically significant.

There were 327 patients with pleural effusion who had pfADA analyzed and were eligible for the study. Of the patients, 26 had transudative pleural effusions, and 25 were excluded due to incomplete data. The remaining 276 patients were eligible for analysis. Thirty-two patients were diagnosed with definite TPE, 38 with clinical TPE, and 206 with non-TPE ( Figure 1 ). Flow diagram. TPE, tuberculous pleural effusion. Table 1 presents the sociodemographic, clinical, and laboratory characteristics, as well as the associated factors, of patients with exudative pleural effusion. Chronic lung diseases include chronic obstructive pulmonary disease, bronchiectasis, bronchial asthma, interstitial lung disease, and lung malignancy. Data are presented as median (interquartile range), n (%), or mean ± standard deviation. *, statistically significant. † , Mann-Whitney U test; ‡ , Pearson chi-squared; § , Fisher’s exact test; ¶ , independent t-test. ADA, adenosine deaminase; LDH, lactate dehydrogenase; TPE, tuberculous pleural effusion. The cohort of 276 patients in this study presents a broad age range, with a median age of 61 years (IQR, 43–72 years). Most patients were male (65.2%), while females accounted for 34.8%. Racially, the largest group was Malay (54.7%), followed by Chinese (33.3%), Indian (7.2%), and other ethnicities (4.7%). Comorbidities were common, with hypertension (41.7%) and diabetes mellitus (31.9%) being the most prevalent. Other notable comorbidities included dyslipidaemia (21.4%), cardiovascular disease (21.7%), and CKD (19.9%). Less common conditions were chronic lung disease (11.2%), chronic liver disease (6.2%), and haematological malignancies (3.6%). Among the 276 patients, TPE was confirmed in 11.6% (definite TPE), while 13.8% had suspected TPE (clinical TPE), and 74.6% were diagnosed with non-TPE causes. Blood investigations revealed a median white blood cell count of 9.3×10 9 /L (IQR, 7.0–13.0). Serum protein levels were within the normal range at 68 g/L (IQR, 62–74), while serum albumin was slightly lower at 27 g/L (IQR, 21–32), which may indicate chronic disease or malnutrition. LDH levels had a median of 261.5 U/L (IQR, 202.0–350.8), suggesting cellular damage or infection. The pleural fluid analysis demonstrated a wide variation in lymphocyte counts, with a significantly elevated median of 432.60 cells/mL (IQR, 124.75–1,148.65). The median pleural fluid protein level was 45 g/L (IQR, 32.3–53.0), and LDH levels were significantly elevated, with a median of 432 U/L (IQR, 243.0–945.1), consistent with the characteristics of exudative effusions. The median pfADA level was 20.76 U/L (IQR, 11.03–51.00), which is often used to diagnose TPE, with levels above 40 U/L typically indicating TB. The LDH/ADA ratio had a median of 21.07 (IQR, 10.00–44.66), which can help distinguish TPE from other causes of pleural effusion. When comparing TPE patients with non-TPE patients, several significant differences emerged. TPE patients were notably younger, with a median age of 45 compared to 64 years for non-TPE patients (P<0.001). Comorbidities such as hypertension, cardiovascular disease, and CKD were more prevalent in non-TPE patients (P=0.044, P=0.04, P=0.04, respectively). Blood investigations revealed that TPE patients had lower total white blood cell counts (P<0.001) but higher protein (P=0.001) and albumin (P=0.03) levels. Pleural fluid analysis also showed higher protein (P<0.001) and ADA (P<0.001) levels in TPE patients, with a significantly lower LDH/ADA ratio (P<0.001) compared to non-TPE patients. In contrast, no significant differences were found in gender or race between the TPE and non-TPE groups. Smoking history was similar in both groups, and specific comorbidities, such as diabetes mellitus, dyslipidaemia, and chronic lung disease, did not show significant variation. Additionally, pleural fluid lymphocyte counts and LDH levels did not show significant differences between the two groups. These findings suggest that certain blood and pleural fluid markers, as well as the patient’s age, are valuable in differentiating between TPE and non-TPE causes of pleural effusion. The most common cause of exudative pleural effusion was non-TPE causes (n=206, 74.6%) compared to TPE (n=70, 25.4%) ( Table 2 ). The leading cause of exudative pleural effusion was PPE (n=116, 42.0%). The PPE patients were primarily diagnosed based on bacteriologically confirmed pleural fluid culture results in 18 patients, while the remaining patients were diagnosed through clinical, biochemical, and imaging findings. MPE, malignant pleural effusion; NSP, non-specific pleuritis; PPE, parapneumonic effusion; SLE, systemic lupus erythematosus; TPE, tuberculous pleural effusion. TPE was the second most common diagnosis in this cohort (n=70, 25.4%); 32 patients with a definite TPE diagnosis were confirmed by positive Mycobacterium tuberculosis (MTB) culture in pleural fluid, pleural fluid MTB polymerase chain reaction (PCR), or histologically confirmed via pleural biopsy. An additional 38 patients were primarily diagnosed through the decisions of respiratory physicians, based on clinical symptoms, radiologic imaging, thoracoscopy findings suggestive of TPE, or clinical improvement after empirical treatment for TB. Malignant pleural effusion was diagnosed in 60 patients (21.7%) among 276 patients. The origin of the malignancies includes lung (n=41, 14.9%), gynaecological (n=4, 1.4%), haematological (n=3, 1.1%), and hepatobiliary (n=2, 0.7%) origins. The remaining malignancies, originating from melanoma, angiosarcoma, prostate, and gastric cancer, each had one patient. Six patients had unknown origins, which required further detailed investigations. Most patients were histologically diagnosed through pleural fluid cytology and biopsy. Non-specific pleuritis, also known as fibrinous pleuritis, is a general term describing chronic pleural inflammation, as reported in 19 patients (6.9%). A relatively less common cause of exudative pleural effusion was uremic pleurisy (n=7, 2.5%). Other rare causes include amyloidosis (n=2), systemic lupus erythematosus-related pleural effusion, and thoracic endometriosis, which had one patient each. Table 3 presents a multivariate analysis identifying independent risk factors for TPE in patients with exudative pleural effusion. The odds ratio (OR), confidence interval (CI), and P values help determine which factors are significantly associated with TPE. *, statistically significant. ADA, adenosine deaminase; CI, confidence interval; LDH, lactate dehydrogenase; TPE, tuberculous pleural effusion. A younger age is significantly associated with TPE (OR =0.980, P=0.043). In addition, higher pleural fluid protein (OR =1.066, P=0.001) and higher pleural ADA (OR =1.008, P=0.01) are strong predictors of TPE. A lower LDH/ADA ratio (OR =0.951, P=0.001) is significantly associated with TPE. In contrast, comorbidities like hypertension, cardiovascular disease, and CKD do not independently predict TPE. While they were more common in non-TPE patients in univariate analysis ( Table 2 ), this association was not significant in multivariate analysis, suggesting that other factors likely confounded it. The ROC curve analysis of pfADA for predicting TPE showed excellent diagnostic performance [area under the curve (AUC) =0.852, P<0.001], with the optimal cutoff value of ADA ≥36 U/L, yielding a sensitivity of 82.9% and a specificity of 80.6% ( Figure 2 ). ROC curve analysis for pfADA to predict TPE in patients with exudative pleural effusion. AUC, area under the curve; CI, confidence interval; pfADA, pleural fluid adenosine deaminase; ROC, receiver operating characteristic; TPE, tuberculous pleural effusion. Figure 3 illustrates the ROC curve analysis for the pleural fluid LDH/ADA ratio in predicting TPE, with an AUC of 0.795 (95% CI: 0.737–0.852, P<0.001), indicating moderate diagnostic accuracy, and the optimal cutoff value of ≤16. It has a sensitivity of 72.9% and a specificity of 70.9%, which are lower than those of ADA alone (sensitivity =82.9%, specificity =80.6%). ROC curve analysis for pleural fluid LDH/ADA ratio to predict TPE in patients with exudative pleural effusion. AUC, area under the curve; CI, confidence interval; LDH/ADA, lactate dehydrogenase/adenosine deaminase; ROC, receiver operating characteristic; TPE, tuberculous pleural effusion. Table 4 illustrates the performance of pfADA in diagnosing TPE. The highest sensitivity and negative predictive value (NPV) are observed with pfADA ≥36 U/L (82.86% and 93.22%, respectively), making it the best option for diagnosing TPE and ruling out TPE when the test is negative. However, due to the moderate positive predictive value (PPV) (58.59%), some false positives may occur, indicating that other diagnostic tests should be used in conjunction with pfADA to confirm TPE. ADA, adenosine deaminase; CI, confidence interval; LDH, lactate dehydrogenase; NPV, negative predictive value; pfADA, pleural fluid adenosine deaminase; PPV, positive predictive value; TPE, tuberculous pleural effusion. PfADA ≥40 U/L had higher specificity than pfADA ≥36 U/L (83.01% vs. 80.60%), meaning it was better to rule out false TPE. PfADA ≥40 had a high NPV (90%), overall making it good at ruling out TPE cases. In contrast, moderate sensitivity (72.86%) and PPV (59.30%) mean that while many TPE cases will test positive, some will be missed, and false positives can occur. A pleural fluid LDH/ADA ratio ≤16 alone had lower specificity (70.87%), NPV (88.48%), PPV (45.95%), and overall accuracy (71.38%), making it less effective in diagnosing TPE cases compared to ADA-based markers. This test should not be used alone for diagnosing TPE, but can be a supplementary tool alongside pfADA levels, microbiology, and imaging findings. Combining both pleural fluid LDH/ADA ratio ≤16 and ADA ≥36 U/L had the highest specificity of 90.78%, PPV of 69.84%, and overall accuracy of 83.70%. This combined approach may provide the most balanced diagnostic performance. It was the best way to ensure a precise diagnosis and confirm TPE when positive.

ADA is a valuable supplementary diagnostic tool for TPE because its levels are typically elevated in response to the immune system’s reaction to Mycobacterium tuberculosis . ADA levels greater than 40 U/L strongly indicate TPE, especially in settings where microbiological confirmation through cultures or molecular testing may be delayed or unavailable, thereby improving the sensitivity of diagnosis in resource-limited areas. In TB-endemic areas like ours, pfADA levels are critical in diagnosing TPE. Our study identified ADA ≥36 U/L as the optimal threshold for TPE diagnosis, providing higher sensitivity (82.86%) and NPV (93.22%) compared to the traditional cutoff of ADA ≥40 U/L. Although the numerical difference between these thresholds appears modest, the clinical significance lies in the substantial improvement in sensitivity and NPV, which are crucial for excluding TPE and minimizing missed diagnoses in TB-endemic settings. Using ADA ≥36 U/L improves case detection and enhances clinical confidence in ruling out TPE when results are negative, with an overall diagnostic accuracy of 80.80%. In contrast, ADA ≥40 U/L offers slightly higher specificity (83.01%), reducing false-positive diagnoses, but at the cost of lower sensitivity (72.86%), potentially leading to missed TPE cases. Given the high morbidity and public health implications of delayed or missed TPE diagnosis, prioritizing sensitivity and NPV is clinically justified in endemic settings. Similar observations supporting locally optimized lower ADA cutoffs have been reported in regional cohorts, including a large Hong Kong study that identified a diagnostic cutoff of ~26.5 U/L using ROC analysis, reinforcing the importance of validating region-specific thresholds rather than universally applying a single cutoff value ( 21 ). Accordingly, ADA ≥36 U/L is recommended as the preferred screening threshold for diagnosing TPE in our setting. When comparing pfADA with the LDH/ADA ratio, pfADA alone is a better diagnostic marker for predicting TPE. Our study established a local cutoff value for the LDH/ADA ratio of ≤16, which demonstrated lower sensitivity (72.86%) and specificity (70.87%), making it less effective in both detecting actual TPE cases and excluding non-TPE cases. While the ratio exhibited a high NPV (88.48%), suggesting it is better for ruling out TPE, its low PPV (45.95%) limits its reliability as a standalone diagnostic test. With an overall accuracy of 71.38%, the LDH/ADA ratio performs less favorably than ADA-based thresholds. Nevertheless, it may provide incremental diagnostic value when used in conjunction with pfADA, particularly in borderline or diagnostically uncertain cases, alongside microbiological tests, imaging, or pleural biopsy. We found that younger patients with exudative pleural effusion, elevated pfADA (median pfADA in TPE: 55.25 U/L), and high pleural protein levels are more likely to have TPE. Pleural fluid ADA emerged as one of the strongest indicators of TPE. At the same time, older patients with comorbidities such as hypertension, cardiovascular disease, and CKD were more likely to have non-TPE effusions, often due to malignancy or infection. Although LDH alone was not very useful in differentiating TPE from non-TPE, the pleural fluid LDH/ADA ratio showed potential as a secondary marker for diagnosing TPE. While pleural lymphocyte counts tended to be higher in TPE cases, they were not significantly different in this study. Further multivariate analysis identified independent risk factors for TPE in patients with exudative pleural effusion. Younger age was significantly associated with TPE (OR =0.980, P=0.043), suggesting that the likelihood of TPE decreases as age increases. This finding aligns with the higher incidence of TB in younger individuals and the prevalence of non-TPE effusions, such as malignancy or heart failure, in older adults. Higher pleural fluid protein and ADA levels were significantly associated with TPE, reinforcing the role of ADA as a key diagnostic marker for TPE. Additionally, a lower LDH/ADA ratio (OR =0.951, P=0.001) was significantly associated with TPE, indicating that ADA levels rise disproportionately compared to LDH in TB. Interestingly, comorbidities such as hypertension, cardiovascular disease, and CKD did not independently predict TPE in multivariate analysis, suggesting that other factors may have confounded their association with non-TPE in univariate analysis. It is acknowledged that the clinical TPE group represents a heterogeneous population with variable diagnostic certainty. However, this heterogeneity reflects real-world diagnostic challenges in TB-endemic settings, where microbiological or histopathological confirmation is frequently unattainable due to the paucibacillary nature of pleural TB. To address this limitation, a composite reference standard incorporating clinical, radiological, microbiological, and histopathological criteria was employed, consistent with prior high-quality studies on TPE. The limitations of pfADA testing include the potential for false positives in conditions such as PPE, lymphoma, and empyema, which may result in falsely elevated ADA levels. This can lead to the premature initiation of anti-TB treatment, which may expose patients to adverse reactions and complications from the drugs. False negatives can occur, particularly in immunocompromised patients (e.g., those with HIV co-infection), where pfADA levels may be lower than expected. Additionally, pfADA isoenzyme testing was not performed in this study. ADA isoenzyme 2, in particular, is known to be elevated in TB and could serve as a differentiating factor between TB and non-TB causes of pleural effusion. However, this test is not widely available or commonly performed in clinical settings ( 27 , 28 ). Several additional limitations should be acknowledged. First, the retrospective design may introduce selection and information bias, as data completeness and diagnostic investigations were dependent on existing medical records and routine clinical practice. Second, the inclusion of a “clinical TPE” group, defined partly by physician judgment and clinical response to anti-TB therapy, may have introduced incorporation and verification bias, as diagnostic decisions could have been influenced by pleural fluid ADA levels and other supportive findings. Third, although a composite reference standard was used to reflect real-world practice in TB-endemic settings, the heterogeneity of the clinical TPE group may have affected the precision of diagnostic accuracy estimates. Fourth, pleural fluid ADA measurements were performed at different laboratories using distinct assay platforms, which may have introduced inter-laboratory variability despite adherence to standardized protocols. Variations in laboratory methodologies, calibration of spectrophotometric instruments, and differences in quality control measures may introduce inconsistencies in ADA level measurements. While both laboratories follow standardized protocols, inter-laboratory differences may affect the uniformity of results, potentially influencing the sensitivity and specificity of pfADA in diagnosing TPE. However, the guidelines do not mention which lab method to use and only provide a value as the cutoff. This variability further supports the rationale for establishing locally validated ADA cutoff values rather than relying on universal thresholds. Fifth, subgroup analyses restricted to microbiologically or histologically confirmed TPE were limited by sample size and may have reduced statistical power to detect differences across diagnostic thresholds. Finally, stratified analyses comparing specific non-TPE etiologies (e.g., parapneumonic versus malignant pleural effusions) were not the primary focus of this study and may warrant further investigation in future prospective studies. This study offers several strengths that enhance its reliability and clinical relevance. The large sample size of 276 cases increases statistical power, while data collected from three centers improves generalizability and reduces institutional bias. A comparative analysis of the pleural fluid LDH/ADA ratio optimizes the diagnostic yield for TPE. The study also addresses regional variations in diagnostic performance by establishing an optimal local ADA threshold. Rigorous statistical analysis, including the ROC curve and Youden Index, refines diagnostic thresholds, and the precise patient classification minimizes misclassification bias, ensuring accurate subgroup analysis. Recent studies have demonstrated that integrating pfADA with additional clinical and biochemical parameters improves the diagnostic discrimination of TPE in patients with unexplained pleural effusions. In particular, prediction models and diagnostic flowcharts incorporating ADA have been shown to enhance specificity and support clinical decision-making, especially in patients with elevated ADA levels ( 29 , 30 ). Integrating pfADA with the LDH/ADA ratio provides a more refined, clinically meaningful diagnostic approach for TPE. Our hypothesis-generating stepwise diagnostic algorithm for TPE ( Figure 4 ) is designed to maximize sensitivity for early identification of TPE while selectively improving specificity in borderline or equivocal cases, thereby reducing both missed diagnoses and unnecessary treatment. While prospective and external validation are required, this model provides a biologically plausible and clinically practical framework that may support diagnostic decision-making in TB-endemic settings. Hypothesis-generating stepwise diagnostic algorithm for TPE integrating pfADA and the LDH/ADA ratio. LDH/ADA, lactate dehydrogenase/adenosine deaminase; pfADA, pleural fluid adenosine deaminase; TPE, tuberculous pleural effusion. The diagnostic model ( Figure 4 ) optimizes the diagnosis of TPE by incorporating a two-step approach based on pfADA levels and the LDH/ADA ratio. A locally established ADA cutoff of ≥36 U/L as the primary screening test provides high sensitivity, ensuring the early identification of TPE cases. For borderline cases, applying the pleural fluid LDH/ADA ratio (cutoff 16) further refines diagnostic accuracy, distinguishing moderate from low-probability cases. The expected benefits of this model include higher specificity (90.78%) by reducing false positives and increased reliability for borderline cases where pfADA alone is inconclusive. If both tests are inconclusive, further tests such as microbiology, interferon-gamma, GeneXpert MTB/RIF assay, pleural biopsy, and imaging findings should be considered before deciding on the diagnosis of TPE. This stepwise approach enhances diagnostic confidence, minimizing misclassification and improving clinical decision-making. The model could be further enhanced by incorporating the predominant cell type (e.g., lymphocytosis). However, this data is unavailable in our study, as we only have the absolute lymphocyte cell count. Future studies should employ prospective, multicenter designs with standardized diagnostic protocols and stratified comparisons between tuberculous, parapneumonic, and malignant pleural effusions to enhance biological and clinical interpretability. Sensitivity analyses restricted to microbiologically or histopathologically confirmed TPE, along with inclusion of larger numbers of definite TPE cases, would strengthen diagnostic accuracy estimates and enable more precise validation of optimal cutoff values. In addition, evaluation of pleural fluid LDH isoenzymes and ADA isoenzymes may provide greater pathophysiological specificity than total enzyme levels and further refine diagnostic models for TPE across different epidemiological settings.

Establishing a new local cutoff value for pfADA is crucial for enhancing diagnostic accuracy and informing clinical decision-making in regions with high TB prevalence, thereby potentially improving patient outcomes and resource allocation. PfADA is an excellent marker for diagnosing TPE, with a threshold of ≥36 U/L providing improved sensitivity (82.9%) and specificity (80.6%) compared to the conventional ≥40 U/L cutoff (sensitivity, 72.9%; specificity, 83.0%). Combining both pleural fluid LDH/ADA ratio ≤16 and ADA ≥36 U/L resulted in the highest specificity, PPV, and overall accuracy. Additional biomarkers and diagnostic tests are crucial when ADA and LDH/ADA results are inconclusive, as they ensure an accurate diagnosis and guide effective treatment decisions.

The article’s supplementary files as