Diagnostic performances and safety of endobronchial ultrasound-guided transbronchial needle aspiration in the older patients: a systematic review and meta-analysis.

Introduction
Lung cancer is a common cause of cancer-related death worldwide [1]. The incidence of lung cancer highly affects the older population, with a median diagnosis age of 70 years, and approximately two-thirds of the cases are observed in individuals aged 65 years or older [2]. Prognosis is influenced by various tumor- and patient-related factors, including stage, performance status, and comorbidities [3]. The development of tyrosine kinase and immune checkpoint inhibitors that specifically target driver mutations has shown promising therapeutic outcomes. These advances suggest improved survival prospects, even among older patients with lung cancer [4]. Therefore, the precise diagnosis and staging of lung cancer are crucial for determining the optimal treatment approach.
Accurate diagnosis of mediastinal diseases requires optimal tissue sampling, preferably guided by computed tomography or positron emission tomography, when mediastinal lesions are identified [5]. Traditionally, the investigation of mediastinal masses or lymphadenopathy relied on surgical procedures, such as mediastinoscopy, video-assisted thoracic surgery, and anterior left mediastinotomy. Since 2010, endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) has become the standard minimally invasive technique for mediastinal staging, effectively replacing mediastinoscopy as the preferred approach for mediastinal exploration [5] This technique provides several advantages, including a high level of diagnostic accuracy, decreased reliance on more invasive procedures such as mediastinoscopy, and an overall favorable safety profile [5].
Due to the high prevalence of pulmonary and cardiac comorbidities in older patients, a thorough risk–benefit assessment is essential prior to performing bronchoscopic procedures. Data from the National Lung Cancer Audit in the UK revealed that patients over the age of 70 years had a reduced likelihood of receiving both a histological diagnosis and cancer treatment, despite adjustments for stage, performance status, and comorbidity [6]. Considering that comorbidities and impaired performance status are important determinants of procedural risk, the administration of optimal sedation and anesthesia may be limited, potentially shortening examination time and reducing sampling adequacy [7]. Furthermore, as the prevalence of malignant disease increases with age, the diagnostic performance of flexible bronchoscopy in older patients may differ from that in younger cohorts [2]. The purpose of this systematic review and meta-analysis was to examine the clinical efficacy and safety of EBUS-TBNA in older patients compared with younger patients.

Methods

Data sources and search strategy
A comprehensive review of the scientific literature was conducted to identify relevant articles published prior to April, 2024, utilizing three electronic databases: PubMed, Embase, and Cochrane Central Register. The following keywords were used to search for articles: (EBUS or EBUS-TBNA or endobronchial ultrasound or endobronchial ultrasonography) and (lung or pulmonary or mediastinal or lymphadenopathy or lymph node) and (older or old or age or aging). This systematic review adhered to the guidelines outlined in the Preferred Reporting Items for Systematic Reviews and Meta-analyses Statement [8]. The study protocol was registered in PROSPERO, an international database for the prospective registration of systematic reviews. This can be identified using registration number CRD42024535934. Given that this investigation involved a systematic review of the existing literature, informed consent or ethical approval was not required.

Inclusion criteria
Our systematic review and meta-analysis included studies that met the following criteria: (1) randomized controlled trials or observational studies examining the use of EBUS-TBNA, (2) focused on older patients (defined as those aged ≥ 65 or ≥ 70 years, depending on the criteria used in each study) for lung cancer diagnosis or staging, and (3) reported data on the diagnostic yield of the procedure. We restricted our search to full-length studies or letters published in peer-reviewed English journals. Review articles, abstracts, case reports, editorials, and extension or post-hoc trials were excluded.

Data extraction, and bias assessment
The authors independently reviewed the studies that met the predefined eligibility criteria based on their titles and abstracts. Subsequently, a comprehensive examination of the full texts was conducted to identify potentially eligible studies. References from relevant articles were manually investigated to identify additional pertinent data. The following information was extracted from each study’s dataset: author, publication year, trial design, study country, study type, total number of subjects, performance status, type of sedation, dose of midazolam, type of EBUS bronchoscope, needle size, procedure time, lymph node size, rate of malignancy established using EBUS-TBNA, and diagnostic yield.
Study quality was assessed by two reviewers using the Newcastle-Ottawa Scale, a tool used to evaluate the quality of non-randomized studies [9]. The scale assesses each study based on eight items grouped into three domains: selection of study participants, comparability of study groups, and ascertainment of either exposure or outcome of interest. Based on the total score across all items, studies scoring 6 points or higher were considered to have a low risk of bias. Any discrepancies identified during the study selection, data extraction, or bias assessment phases were resolved through collaborative discussion.

EBUS-TBNA procedure
All procedures were performed using a linear-array convex ultrasound bronchoscope under conscious sedation or general anesthesia, allowing clinicians to visualize lymph nodes and lesions adjacent to the airways. Topical anesthesia with lidocaine was applied to the upper airway prior to bronchoscope insertion. Real-time ultrasound imaging was used to identify mediastinal and hilar lymph nodes as well as peribronchial lesions. Under direct visualization, a dedicated 21- or 22-gauge aspiration needle was advanced through the working channel of the bronchoscope and introduced across the bronchial wall into the target lymph nodes or lesions. Typically, 2–4 passes were made per target to ensure adequate sampling. Aspirated material was processed for cytological examination, histopathological evaluation, and, when applicable, molecular testing. All patients were monitored continuously for oxygen saturation, heart rate, and blood pressure throughout the procedure, and any adverse events were recorded [5].

Data synthesis and statistical analysis
A proportional meta-analysis was conducted to calculate the pooled diagnostic yield of EBUS-TBNA in both the older and younger patients. Diagnostic yield was determined by dividing the number of patients with positive diagnoses by the total number of cases. The pooled proportions and 95% confidence intervals (CIs) for the individual studies were also calculated. The primary outcome was the diagnostic yield of the EBUS-TBNA. The diagnostic criteria for the overall diagnosis included both malignant and benign diseases. The criteria for malignancy included either histological confirmation of malignancy during mediastinal staging or a clinical diagnosis of lung cancer made by the responsible physician. And we conducted age-stratified subgroup analysis in older group.
Inter-study statistical heterogeneity was evaluated using I2 statistics, which ranged from 0% to 100% [10]. If I2 exceeded 50%, indicating substantial between-study heterogeneity, a random-effects model was used; otherwise, a fixed-effects model was used. [10] Statistical significance was determined by a P value of less than 0.05. Statistical analyses were performed using Stata statistical software (Version 14.2, Stata Corp LP, College Station, TX, USA).

Results

Study search
Figure 1 illustrates the selection process using a flow diagram. Initially, 2,454 records were identified. Following the removal of duplicates, 2,113 articles were considered eligible for abstract review based on their titles. Subsequently, 12 articles were subjected to full-text review. Among these, six records were excluded based on the criteria outlined in Fig. 1. Finally, six articles meeting the defined inclusion criteria were included [11–16].

Table 1 presents an overview of the baseline characteristics of the studies analyzed in this systematic review. The analysis involved 5,206 patients, with 1,724 classified in the older group and 3,482 in the younger group. Studies included in this meta-analysis were published between 2013 and 2022. The number of subjects in each trial varied from 109 to 2,444. Of the studies included, four performed EBUS procedures using Olympus bronchoscopes [12–15], whereas Fujifilm and Pentax devices were used in each study [11, 16].

With respect to age criteria, three studies defined older patients as those aged ≥ 65 years [11, 12, 14], whereas the other three studies used 70 years as the cutoff [13, 15, 16]. Based on the Newcastle-Ottawa Scale assessment, all included studies scored 7 points or higher, indicating a low risk of bias.

Diagnostic performance and accuracy of EBUS-TBNA in the older and the younger group
Figure 2 illustrates forest plots for the diagnostic yield of EBUS-TBNA. Based on the random-effects model, the pooled diagnostic yield for overall diagnosis was 0.96 (95% CI, 0.94 to 0.99) in older patients and 0.97 (95% CI, 0.95 to 0.99) in younger patients (Fig. 2A and B). The yield was similar between the two groups (P = 0.81). Among older patients, the diagnostic yield reported across studies ranged from 0.79 to 1.00, whereas in younger patients, it ranged from 0.71 0.99. Substantial heterogeneity was observed, with I² values of 80.76% and 86.92% in the older and younger patients, respectively. In the age-stratified subgroup analysis, the pooled diagnostic yield of EBUS-TBNA for malignancy was 0.97 (95% CI, 0.96 to0.98) with a 65-year age cutoff and 0.90 (95% CI, 0.82 to 0.98) with a 70-year cutoff (P < 0.01).

Figure 3 presents the forest plots for the diagnostic yield of EBUS-TBNA for detecting malignancy. In the random-effects model, the pooled diagnostic yield for malignancy detection was 0.59 (95% CI, 0.47 to 0.71) in older patients and 0.51 (95% CI, 0.28 to 0.73) in younger patients (Fig. 3A and B). This yield was significantly higher in the older patients than in the younger patients (P < 0.01). The diagnostic yield ranged from 0.47 to 0.82 in older patients and from 0.21 to 0.93 in younger patients. Considerable heterogeneity was noted, with I² values of 94.48% for older patients and 99.48% for younger patients. The age-stratified subgroup analysis revealed that the diagnostic yield for malignancy was significantly higher at the 70-year cutoff (0.67; 95% CI, 0.46 to 0.88) compared to the 65-year cutoff (0.49; 95% CI, 0.45 to 0.63; P < 0.01).

Secondary outcomes
All included trials compared the rate of any complications between the older and younger groups undergoing EBUS-TBNA [11–16]. The pooled estimates using a fixed-effect model with the Mante-Haenszel method demonstrated that the rate of any complications was not significantly different between the two groups (risk ratio, 0.95; 95% CI 0.74 to 1.23; P = 0.72; I2 = 0%, Fig. 4A). The pooled rates of complications were 5.13% and 5.20% in the older and younger groups, respectively.

We retrieved data on the mean dose of midazolam administered to the two groups during the EBUS-TBNA from four trials [12, 13, 15, 16]. The pooled mean doses of midazolam were 2.55 mg and 2.83 mg in the older and younger groups, respectively. In the pooled estimates, the older group received a lower mean dose of midazolam compared to the younger group (mean difference [MD] −1.11 mg; 95% CI −1.58 to −0.63; P < 0.01; I2 = 91.2%, Fig. 4B).
Data regarding the duration of EBUS-TBNA procedures were extracted from four trials [11–13, 15]. The pooled mean procedure duration of EBUS-TBNA was 24.87 min and 22.38 min in the older and younger groups, respectively. The pooled estimates showed that the mean duration of the procedure was similar between the older and younger groups (MD −0.03 min; 95% CI −1.44 to 1.38; P = 0.97; I2 = 63.8%, Fig. 4C).

Discussion
A recent retrospective study found that flexible bronchoscopy in patients aged ≥ 80 years yielded diagnostic and safety outcomes comparable to those in younger patients. These findings indicate that flexible bronchoscopy can be feasibly and safely performed in the older population without additional risk of complications. Clinically, this implies that advanced age alone may not necessarily be considered a contraindication for flexible bronchoscopy [17].
Compared with conventional flexible bronchoscopy, EBUS-TBNA further exacerbates concerns about periprocedural complications. The larger outer diameter of the EBUS scopes and longer procedural durations, along with various host-related factors, can present particular challenges in older patients [17]. In the present study, we investigated the diagnostic yield and safety of EBUS-TBNA in older patients using a systematic review and meta-analysis. The pooled diagnostic yield of EBUS-TBNA among the older group was 0.96, which is consistent with the findings observed for EBUS-TBNA in previous studies [18]. And the diagnostic yield of EBUS-TBNA for overall diagnosis was comparable between the older and younger groups.
Our pooled analysis showed that the diagnostic yield for malignant diseases in the older group was 0.59, which was higher than that observed in the younger group. Considering that the incidence of malignant disease increases with age, our findings are consistent with this trend [2]. The present study also revealed that the 70-year cutoff showed a higher malignant detection rate compared to the 65-year cutoff, indicating a trend toward increased diagnostic yield for malignancy with advancing age. Non-small cell lung cancer is frequently linked to better prognosis in patients who have positive driver gene mutations or biomarkers indicative of responsiveness to immune checkpoint inhibitors [4]. A retrospective study of older patients over 10 years reported that the lack of genetic mutations and low programmed death-ligand 1 expression were likely key factors in the decision to provide the best supportive care [19]. Even in older patients, active pathological diagnosis is expected to be beneficial for treatment and prognosis.
We analyzed the data using diagnostic yield rather than diagnostic testing accuracy, such as sensitivity and specificity. This decision was made because some of the included studies did not report two-by-two tables, which prevented the identification of true negative results and calculation of diagnostic accuracy. The high sensitivity of EBUS-TBNA in the older group could be attributed to the low rate of false-negative sampling. In older patients with negative findings for malignancy on EBUS-TBNA, there might be a lower likelihood of undergoing repeat sampling or surgical approaches such as mediastinoscopy to confirm malignancy than in the younger group. One study included our analysis indicated that the older group underwent more extensive clinical and radiological follow-up to establish the final diagnosis than the younger group [12]. However, it is important to note that this follow-up might not substitute for tissue sampling, which could have affected the sensitivity.
Sedation and anesthesia play a crucial role in enhancing the procedural comfort of patients and facilitating procedures. Considering that comorbidities and performance status are primary contributors to potential complications in older patients, there are concerns about administering sedative agents [20]. Previous studies have reported that EBUS-TBNA procedures performed under general anesthesia or deep sedation can entail increased risks for older individuals and often require the presence of an anesthesiologist [21]. Patients enrolled in our study frequently underwent EBUS-TBNA with conscious sedation administered with midazolam without deep sedation or general anesthesia. Although the older group received a lower dose of midazolam and had comparable procedure times to the younger group, they showed similar diagnostic yields and complication rates during EBUS-TBNA procedures. The outcomes of a recent retrospective trial of 280 patients using deep sedation were similar to those of our study [20]. In this trial, propofol was administered as a sedative agent for EBUS-TBNA, with an examination conducted on procedure duration, recovery time, hemodynamic data, and complication rates in patients aged ≥ 65 years compared to those aged 45 years and younger [20]. We hypothesized that the minimal invasiveness of the EBUS-TBNA procedure and utilization of conscious sedation would effectively reduce complication rates. This suggests that if there are no significant concerns regarding respiratory conditions or performance status during EBUS-TBNA procedures, it might indicate the feasibility of performing the procedure safely irrespective of age. Additionally, we identified a retrospective observational study to investigate the safety and effectiveness of EBUS-TBNA in very old patients aged 80 years and older [19]. The study’s findings, which indicated an 89% diagnostic rate for malignant diseases and a 5% complication rate, were consistent with our own results [19].
While the terms “older” and “younger” are used for categorical analysis in this study, they represent practical approximations rather than strict biological classifications. In clinical practice, eligibility for EBUS-TBNA is determined not solely by age but by a combination of patient preference, procedural indication, and physiological capacity to undergo both diagnostic bronchoscopy and subsequent oncologic management. Therefore, the “older” group in this analysis refers to patients aged ≥ 65 or ≥ 70 years who are considered clinically fit and willing to undergo bronchoscopy. In addition, this selection process may introduce clinician-driven selection bias, as patients deemed unsuitable for bronchoscopy due to frailty, comorbidities, or limited therapeutic options are less likely to be referred for the procedure. And a large U.K. study demonstrated that patients over 70 years of age were less likely to receive both a histological diagnosis and subsequent cancer treatment, even after adjustments for stage, performance status, and comorbidity (6). These findings suggest that the diagnostic rates observed in older patients may be influenced not by the diagnostic yield of EBUS-TBNA itself, but rather by a decreased likelihood of undergoing the procedure, thereby reflecting a potential selection bias. This potential bias should be considered when interpreting the generalizability of our findings to the broader older population.
The current study has some limitations. First, because the present study relied on a small sample size, generalizing our conclusions might be challenging. The restricted number of included trials prevented us from evaluating this bias using a funnel plot for the outcomes. There might have been inherent publication bias, a common issue in meta-analyses, which could have affected our findings. Second, considerable between-study heterogeneity was observed in the diagnostic yield of EBUS-TBNA. Various factors, such as performance status, comorbidities, sarcopenia, and nutritional status could potentially impact the outcomes and safety of the procedure. Nonetheless, because of the scarcity of data in the included studies, a comprehensive investigation of these factors, such as subgroup or meta-regression analyses, was not possible. Third, although complications were prespecified in all included studies, four of the six identified them retrospectively through chart review, which may have led to an underestimation of the actual complication rates.

Conclusions
Our findings suggest that the diagnostic yield and accuracy of EBUS-TBNA in older patients are comparable to those observed in younger cohorts, with no significant increase in complication rates. Although older patients tended to receive lower doses of midazolam, the procedure duration was similar between age groups, supporting the overall feasibility and safety of EBUS-TBNA in this population. These results imply that advanced age by itself is unlikely to represent a limitation when considering EBUS-TBNA for diagnostic purposes.

Supplementary Information