Influence of age on the diagnostic accuracy of soluble biomarkers for tuberculous pleural effusion: a post hoc analysis

Tuberculosis is a serious public health problem globally but especially in developing countries. Tuberculous pleural effusion (TPE) ranks among the most regular forms of extrapulmonary tuberculosis [1]. The definitive diagnosis of TPE relies upon the presence of Mycobacterium tuberculosis in the sputum, pleural effusion, or pleural biopsy specimens [2]. To achieve the gold standard, conventional diagnostic tests include microscopic examination and/or culturing of pleural fluid, sputum, or pleural tissue for acid-fast bacilli or the histopathological demonstration of caseating granulomas in the pleura along with acid-fast bacilli. These tests have recognized limitations in clinical practice, such as low sensitivity, lengthy delay, or invasiveness [3, 4]; as a result, multiple biomarkers in pleural effusion have been investigated, including adenosine deaminase (ADA) and interferon-gamma (IFN-γ) [5].

Differences in the diagnostic value of the same biomarker in the literature is associated with a variety of factors. One such factor is age, which has been reported to be correlated with pleural ADA levels and affects the diagnostic accuracy of TPE [6‐11]. One possible explanation is that elderly people suffer a degradation of immune status with aging, believed to result from a functional decrease in lymphocyte and macrophage levels [10], and this may affect the levels of pleural biomarkers. However, it is still less clear as to whether combining pleural biomarkers with patients’ age could be a valuable method for estimating the probability of TPE.

We conducted two prospective studies and one meta-analysis, finding that the diagnostic value of interleukin 27 (IL-27) is comparable to that of IFN-γ and more accurate than that of ADA [12]. Through a post-hoc analysis of the two independent prospective blind studies, we focused on the influence of age, as an independent covariate, on the diagnostic accuracy of ADA, IFN-γ, and IL-27 in pleural fluid to differentiate between TPE and non-TPE cases.

Diagnosing pleural effusion remains a challenge in clinical practice. Considering different etiologies, tuberculosis is the most common cause of pleural effusion in developing countries [18, 19]. Unfortunately, conventional methods of microscopic examination, culture, or histopathological demonstration for diagnosing TPE have shown low sensitivity or are expensive and invasive [2, 20]. A number of pleural soluble biomarkers have been evaluated [5, 12, 19], yet a problem of insufficient overall diagnostic accuracy has consistently been encountered.

Several studies have introduced predictive models based on clinical parameters and the chemistry profile in pleural fluid to facilitate differential diagnosis between TPE and non-TPE [21], but few studies have looked into the factors affecting pleural biomarkers, especially in relation to age [6‐11]. The majority of published studies are retrospective single-center ones and lack validation. In this post-hoc analysis of two independent prospective blinded studies, we evaluated the diagnostic values of ADA, IFN-γ, and IL-27 in pleural fluid for TPE according to different ages. Our results showed that the levels of these biomarkers were not different between young patients and older patients. In young patients, ADA, IFN-γ, and IL-27 accurately diagnosed TPE, while in older patients, IL-27 and IFN-γ appeared as excellent options to discriminate TPE from non-TPE. Age affected the diagnostic accuracy of pleural ADA for TPE, with a trend towards a reduction in older patients. Of note, this is, to the best of our knowledge, the first study to examine factors affecting pleural IL-27.

Several studies have demonstrated a negative correlation between pleural fluid ADA activity and age in the entire cohort and/or in a subgroup of patients with TPE [6‐11], as age-associated immune decline is likely to affect ADA level. However, we did not find any difference between young patients and older patients regarding the concentrations of ADA, IFN-γ, and IL-27. This discrepancy may be partly related to the study inclusion criteria, the number of participants, and variations in age groups. Thus, the true role of age-associated changes in pleural biomarker levels and the precise mechanisms remains an important area for future inquiry/investigation.

Several previous studies have reported relatively high diagnostic accuracy for pleural ADA, IFN-γ, and IL-27 for TPE [12, 22, 23]. One published study suggested that, in a region with a high prevalence of tuberculosis, the routine pleural fluid measurement of ADA concentration among young patients displays a reliable level of diagnostic accuracy [24]. Our research confirmed these earlier findings and further supports the use of ADA, IFN-γ, and IL-27 as excellent diagnostic biomarkers to discriminate TPE from non-TPE in young patients with high sensitivity and specificity values of 100.0 and 100.0%, respectively. Each of these three biomarkers can be used as a rule-in test for TPE when they surpass their threshold values. TPE patients had a lower mean age, and tuberculosis was reported as the most common cause of pleural effusion in patients under 40 years of age in areas with a high incidence of tuberculosis [16], so the higher diagnostic accuracy of these biomarkers may partly be due to the increased prevalence of TPE. Some authors have also reported high ADA levels among patients with other causes of pleural effusion (e.g., pneumonia, empyema, lymphomas, adenocarcinomas, and systemic lupus erythematosus), but all of our young patients with these conditions showed pleural ADA concentrations below our diagnostic threshold. Most TPE cases could be diagnosed by medical thoracoscopy [25]; however, not everyone is a candidate for or willing to undergo thoracoscopy for its invasiveness. Since the detection of ADA is simpler and cheaper, the present study recommends it as the first choice for the diagnosis of TPE in young patients [26], may help to avoid further invasive biopsies in some cases.

IL-27, a member of the IL-12 cytokine family produced by activated antigen-presenting cells, has been reported to control the development of regulatory T cells and IL-17–producing CD4+ T cells [27]. Our previous study showed that the diagnostic performance of IL-27 was comparable to that of IFN-γ and more accurate than that of ADA in the diagnosis of TPE in the total patient population [12]. In the present study, we once again observed that, in both cohorts, IL-27 and IFN-γ constitute excellent biomarkers to diagnose TPE in older patients and their diagnostic values are not affected by patient age. As the predictive value of pleural fluid ADA level for TPE decreases with increasing age, physicians should be cautious when interpreting the pleural ADA level to diagnose TPE if the patient is older. Some studies have suggested that the diagnostic cutoff value of pleural ADA activity for TPE requires adjustment according to different age groups [6, 8]. Although IFN-γ has an advantage in diagnosing TPE, its daily clinical use is limited due to the high cost and lack of acceptable cutoff value [2, 19]. Due to the excellent diagnostic accuracy of pleural IL-27 for TPE in this study, such represents a potential option in routine clinical practice in older patients.

Based on the findings of our study, in regions of high tuberculosis prevalence, we hypothesize that combining pleural biomarkers with patients’ age should be a simple method for estimating the probability of TPE with high sensitivity and specificity. If patients are younger than 40 years of age, we recommend testing for pleural ADA, while, if patients are older than 40 years of age, then IL-27 should be measured. Due to the high diagnostic accuracy of pleural biomarkers according to different ages for TPE, one might consider skipping unnecessary invasive diagnostic procedures and initiating antituberculous therapy in certain groups of patients, and the study holds a certain meaning for health economics, especially in developing countries.

Several limitations of this study should not be ignored. First, although it provided data from two separate prospective cohorts containing more than 250 patients, the total number of cases was still small, especially in the young non-TPE group, as other causes of pleural effusion in young patients were relatively rare from the epidemiological perspective, and the small study sample may be affected by selection bias. Second, the prevalence of tuberculosis, time of pleural fluid collection, standardization of the detection method, or cutoff point may all influence the study results, and a subgroup analysis controlling for confounders was not conducted. Third, our simple flowchart, consisting of age and pleural biomarkers, is useful and accurate for TPE diagnosis. However, it is just a first step and should not be considered as an alternative to biopsy and culture. More invasive methods or other means should potentially follow to establish a definitive diagnosis of TPE. In addition, further studies are required to confirm the mechanism associated with the biomarkers in pleural fluids.

In conclusion, our study reports the use of pleural biomarkers according to different ages may improve the diagnostic accuracy for TPE. In high tuberculosis prevalence settings, in young patients, ADA, IFN-γ, and IL-27 could all accurately diagnosis TPE. In older patients, IL-27 and IFN-γ are excellent biomarkers to discriminate TPE from non-TPE. However, additional studies are needed to provide more convincing data to establish the definite diagnosis of pleural effusion for clinical practice, especially before the results of biopsy or culture become available.