Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is a minimally invasive diagnostic procedure for sampling and evaluation of mediastinal and hilar lymph nodes.1 EBUS-TBNA has been shown to be superior to conventional TBNA and at least equivalent to mediastinoscopy for the evaluation of mediastinal adenopathy in lung cancer staging.2,3 This procedure is safe and highly useful for obtaining cytologic as well as histologic samples that permit adequate studies including immunohistochemistry and driver mutation assessments.4
Traditionally, EBUS-TBNA has been performed using a 22-G needle until 2011, when a larger 21-G needle was introduced. Theoretically, a larger 21-G needle is expected to provide a greater amount of tissue resulting in a higher diagnostic yield. However, prior studies from other organ systems have shown that tissue adequacy may not always correlate with the size of the needle.5–7 Only a few studies on EBUS-TBNA have compared the diagnostic yield and sample adequacy of the newer 21 G needles with the traditional 22 G needles and the results have been inconsistent.1,8–13
Because of these conflicting results, we sought to systematically review the existing data comparing 21 with 22 G aspiration needles.
Literature Search and Data Sources
A systematic search of relevant articles published through September 21, 2014 was performed in MEDLINE (via PubMed), EMBASE, and SCOPUS databases using the following Boolean search strategy: (Endobronchial Ultrasound OR EBUS) AND (Transbronchial Needle Aspiration OR TBNA OR Needle) AND “Gauge.” In addition, bibliographies of the included studies and review articles were hand-searched individually.
Prospective and retrospective observational studies as well as randomized-controlled trials comparing 22 G aspiration needles with 21 G aspiration needles in EBUS-TBNA procedures were included in our study. The study population of interest included all adult patients (above 18 y) who underwent diagnostic EBUS-TBNA in an unselected group of patients. The inclusion criteria included observational and interventional studies with clear comparison arms of EBUS-TBNA with 21 and 22 G needles. We excluded studies on pediatric patients, studies without comparison arms, studies reporting outcomes on patient subpopulations only, and studies published as conference abstracts. The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement was followed.14 Protocol for this meta-analysis was prospectively devised that details the background, objectives, and eligibility criteria of studies, outcomes, and statistical methods. This is available for review upon request to investigators.
Data Extraction and Quality Assessment
Two authors (V.Y. and S.G.) screened and excluded irrelevant studies. Relevant data were extracted by 2 authors (S.G. and R.P.) and checked by another (V.Y.) using a standardized data-extraction table in Microsoft Excel 2010. Additional investigators (M.R.A. and P.K.) participated in the review process when uncertainty about eligibility criteria arose. We obtained data on study characteristics, patient demographic and clinical information, outcome measures including complication rates in the 2 groups.
Data Synthesis and Analysis
The main outcome measures of interest in our study were diagnostic yield, sample adequacy for either histologic or cytologic diagnosis, and the mean number of needle passes. Incidence of procedural complications between the 2 groups was measured as a safety outcome.
Diagnostic yield in our study was defined as the ability to reach a specific diagnosis based on EBUS-TBNA samples. Sample adequacy was defined as (a) the availability of adequate material for cytologic or histologic assessment as defined by the individual studies and (b) the presence of lymphocytes or a definitive diagnosis on cytology. Where available, sample adequacy rates were quantified based on the number of lymph nodes sampled rather than the number of patients in the study. Similarly the mean number of needle passes to obtain tissue samples (with SD) was compared between the 2 study groups.
Data from the studies were compared and interpreted using RevMan 5.3 software (Nordic Cochrane Centre, Cochrane Collaboration, Copenhagen, Denmark, 2014) and Comprehensive Meta-Analysis software (CMA 3.3, Biostat, Englewood, NJ, 2014). Summary odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using the Mantel-Haenszel random-effects method to account for heterogeneity. Study heterogeneity was estimated with I2 statistic, where the value of I2<25% and>75% were defined as low and high heterogeneity, respectively. Publication bias was tested through visual inspection of the asymmetry in funnel plots and further evaluated with Egger’s regression intercept and Begg and Mazumdar rank correlation. A 2-tailed alpha level of 0.05 was used as the level of significance.
A total of 5 eligible studies involving a total of 1720 patients were identified and included in this review (Fig. 1).1,9,11–13 Four studies were excluded from quantitative analysis because of (a) marked variation in reporting of outcomes from the rest of the studies,15 (b) reporting of study results on select patient subgroups only,16 and (c) publication as a conference abstract only.8,10 Of the included studies, 4 were observational studies1,9,11,13 and 1 was randomized clinical trial.12 In 1 study, the same patient cohort served as their controls.11
The study population included in the studies is summarized in Table 1. The mean age was comparable among the 2 groups in these studies. Similarly the prevalence of malignant and benign diseases seemed to be comparable among the 2 groups.
Information regarding the diagnostic yield was available from all 5 studies involving a total of 1720 study participants.1,9,11–13 There was no statistically significant difference in the diagnostic yields between the 21 and the 22 G groups (73.7% vs. 58.5%; OR, 1.04; 95% CI, 0.80-1.35; P=0.80; I2=0%) (Fig. 2).
Information regarding sample adequacy for histologic diagnosis was available from 3 studies involving 1345 participants.1,12,13 There was no statistically significant difference in sample adequacy rates between the 21 and the 22 G group (89.1% vs. 90.0%; OR, 0.94; 95% CI, 0.56-1.59; P=0.82; I2=30%) (Fig. 2). Sample adequacy for cytology alone was only reported in the study by Saji et al13 and was not significant in the 2 groups (100% vs. 96.8%; OR, 2.33; 95% CI, 0.09-59.81; P=0.61) (Fig. 3).
Mean Number of Needle Passes
The mean number of needle passes were reported only by 2 studies and was not significantly different between the 21 and the 22 G group1,13 (mean difference −0.31; 95% CI, −1.10 to 0.47; P=0.44) (Fig. 4). Saji et al13 reported the mean number of needle passes to be 1.9 (SD, 0.4) and 1.8 (SD, 0.7), respectively for the 21 and the 22 G groups. Similarly, Yarmus et al1 reported the mean number of needle passes to be 3.5 (SD, 1.2) and 4.2 (SD, 1.3), respectively. Jeyabalan et al9 reported the mean number of needle passes to be 3.3 and 3.4, respectively, for 21 and 22 G needles. However, no additional information (SD, t statistic, or P-value) was available to permit meta-analysis of this data.
Procedural complications were reported by 4 studies.9,11–13 No major complications such as life-threatening hemorrhage, pneumothorax, pneumomediastinum, or massive hematoma were reported in the included studies. Yarmus and colleagues reported some minor self-limited bleeding from the puncture site. Although, there was no comparative safety data reported in these studies for our analysis, Yarmus et al1 reported that the incidence of these minor complications were not significantly different between the 2 groups.
Study Quality and Publication Bias
Quality assessment of the individual studies was conducted using Quality Assessment of Diagnostic Accuracy Studies-2 checklist (QUADAS-2). The risk of bias was determined to be low among all 4 studies in terms of patient selection and the use of index test. As these studies were primarily designed to study the diagnostic yield rather than the diagnostic accuracy, a reference standard test was not performed. In all studies but one,12 selection of needle gauge for each patient was determined by the performing bronchologist. In 2 studies,9,13 patients with negative diagnostic yields were clinically and radiologically followed up for 6 months to 1 year to ensure true negativity.
On visual examination of the funnel plot, there was some suggestion of publication bias with smaller studies more likely to have a positive effect (Fig. 5). This was further tested with Egger’s regression intercept which was 0.96 (95% CI, −0.44 to −2.35; P=0.11), and Begg and Mazumdar rank correlation (τ=0.70, P=0.08).
One of the included studies by Yarmus et al1 in our meta-analysis was much larger than the remaining 5 studies and might have significantly influenced our results. We performed a sensitivity analysis for diagnostic yield and sample adequacy rates between the 2 study groups by excluding the study by Yarmus and colleagues from our analysis. Sensitivity analysis showed similar results in terms of diagnostic yield with the inclusion of the study by Yarmus and colleagues (OR, 1.04; 95% CI, 0.80-1.35; P=0.80; I2=0%) versus without the study (OR, 1.40; 95% CI, 0.63-3.08; P=0.41; I2=5%) as well as for sample adequacy (89.1% vs. 90.0%; OR, 0.94; 95% CI, 0.56-1.59; P=0.82; I2=30%) versus (OR, 1.48; 95% CI, 0.22-10.07; P=0.69; I2=65%), respectively.
The results of the current meta-analysis suggest that newer 21 G needles during EBUS-TBNA have similar overall efficacy and safety profile as compared with the traditional 22 G needles.
Sampling of mediastinal and hilar lymph nodes is crucial in diagnosing pathologies and staging lung cancers. Currently, there are different techniques available for sampling, including invasive techniques such as mediastinoscopy and video-assisted thoracic surgery, and minimally invasive techniques such as computed tomography–guided fine needle aspiration, conventional TBNA, and EBUS-TBNA. In comparison with mediastinoscopy, EBUS-TBNA has a better safety profile and cost effectiveness along with comparable sensitivity and negative predictive value.17 For lung cancer, EBUS-TBNA has a higher sensitivity, specificity, and diagnostic accuracy in comparison with radiologic staging measures (computed tomography and positron emission tomography).18
Currently, 2 different needle gauges (21 and 22 G) are used for tissue sampling during EBUS-TBNA. The smaller 22-G needle can be easily maneuvered for sampling smaller nodes and lesions at difficult locations; however, the amount of tissue obtained may be limited. In comparison, the recently introduced larger 21-G needle is expected to yield a greater amount of tissue sample and is expected to increase the diagnostic yield.9 These 2 needles have been compared by few observational studies and randomized trials, the results of which have been conflicting.1,8,10–13
On the basis of our literature search, 5 eligible studies were identified comparing the sample adequacy, diagnostic yield, and safety outcomes of the 2 needle types.1,9,11–13 No significant heterogeneity for efficacy outcomes was seen among these studies. Studies by Nakajima et al,11 Oki et al,12 Yarmus et al,1 and Jeyabalan et al,9 showed no overall difference in efficacy outcomes (diagnostic yield, sample adequacy, or mean number of needle passes), whereas Saji et al13 showed a superior performance of 21 G needles as compared with 22 G needles. As the internal diameters of 21 and 22 G EBUS-TBNA needles are equivalent to 20 and 21 G conventional TBNA needles, lack of significant difference in diagnostic yield between the 2 needle types may not be entirely surprising. Also, on subgroup analysis, Jeyabalan and colleagues reported a superior performance of 21 G needles as compared with 22 G needles for characterization of benign lesions and non–small cell lung cancer. In their observational study among 56 patients with mediastinal lesions, Saji and colleagues showed superior sample adequacy and diagnostic yield among patients who underwent sampling with the 21-G needle.13
The rate of procedural complications among the 2 needle types were only reported by 4 studies9,11–13 and involved only minor complications which were similar between the 2 groups. Overall, EBUS-TBNA is a minimally invasive, safe and well-tolerated procedure. Major complications such as pneumomediastinum, pneumothorax, and hemomediastinum occurring during this procedure have been rarely reported in the literature. The use of real-time imaging enables direct visualization of tissues and blood vessels, and hence major vessel puncture rarely occurs during these procedures.
The overall diagnostic yield of the EBUS-TBNA in the 2 needle groups (73.7% in the 21 G group and 58.5% in the 22 G group) are lower than previously reported in the literature.19,20 On closer review of data, this can be largely attributed to the low diagnostic yields reported by Yarmus et al.1 Excluding that study, the cumulative diagnostic yield from the remaining studies are 94.2% and 90.5%, respectively. Yarmus and colleagues report cumulative data from 6 large volume centers of the United States [Beth Israel Deaconess Medical Center, Chicago Chest Center, Johns Hopkins University, Kaiser Permanente (Oregon), Henry Ford Medical Center, and University of Texas MD Anderson Cancer Center], which may be subject to referral bias.1 Also difference regarding the size and location of the mediastinal lesions may have affected the diagnostic yield, which could not be compared across the various studies.
Our study has several limitations. Most of the data in this study come from observational studies and more randomized-controlled trials are anticipated on this topic. Any meta-analysis is subject to publication bias, as studies with a positive result are more likely to be published in the literature. The very small number of studies included in the current meta-analysis makes it difficult to make any conclusive inferences on the summary point estimates. Also the test of heterogeneity is underpowered due to the small number of studies and the chances of occurrence of a beta error is high. The use of random-effects model in the current meta-analysis may have helped minimize this error to some effect. Limited information was available regarding tests of diagnostic accuracy (sensitivity, specificity, positive and negative predictive value) from the selected studies, likely because their methodology did not involve the use of direct surgical sampling in the form of a reference standard test. Hence, we could not perform a cumulative analysis of these measures. Also, we did not compute the diagnostic utility of these 2 needle types among specific subpopulations, for example, benign versus malignant lesions.
In conclusion, the use 21 G needles in EBUS-TBNA appear to be overall comparable with 22 G needles for the evaluation of mediastinal lymphadenopathy in terms of diagnostic yield, sample adequacy, and the number of needle passes required without an increased risk of procedural complications. Whether the diagnostic performance between these 2 needles is similar for specific patient subpopulations remains unclear and needs to be explored by future studies.
1. Yarmus LB, Akulian J, Lechtzin N, et al.. Comparison of 21-gauge and 22-gauge aspiration needle in endobronchial ultrasound-guided transbronchial needle aspiration
: Results of the American College of Chest Physicians Quality Improvement Registry, Education, and Evaluation Registry. Chest. 2013;143:1036–1043.
2. Sharples LD, Jackson C, Wheaton E, et al.. Clinical effectiveness and cost-effectiveness of endobronchial and endoscopic ultrasound relative to surgical staging in potentially resectable lung cancer: results from the ASTER randomised controlled trial. Health Technol Assess. 2012;16:1–75. iii-iv.
3. Almeida FA, Casal RF, Jimenez CA, et al.. Quality gaps and comparative effectiveness in lung cancer staging: the impact of test sequencing on outcomes. Chest. 2013;144:1776–1782.
4. Sharma D, Hovda JT, Dotson TL, et al.. Myths and legends: does needle size affect the diagnostic yield
and accuracy of EBUS-TBNA? Clin Pulm Med. 2014;21:239–240.
5. Franiel T, Fritzsche F, Staack A, et al.. Histopathologic quality of prostate core biopsy specimens: comparison of an MR-compatible biopsy needle and a ferromagnetic biopsy needle used for ultrasound-guided prostate biopsy. Rofo. 2006;178:1212–1218.
6. Gumus M, Cay N, Algin O, et al.. Comparison of 21 and 27 gauge needles for determining sample adequacy
in the aspiration biopsy of thyroid nodules. Diagn Interv Radiol. 2012;18:102–105.
7. Hanbidge AE, Arenson AM, Shaw PA, et al.. Needle size and sample adequacy
in ultrasound-guided biopsy of thyroid nodules. Can Assoc Radiol J. 1995;46:199–201.
8. Grah C, Griff S, Mairinger T, et al.. Comparison of 21 gauge and 22-gauge aspiration needle during endo-bronchial ultrasound-guided transbronchial needle aspiration: a randomised trial. In: Proceedings of the 14th World Conference on Lung Cancer; Jul 3-7, 2011; Amsterdam, Switzerland. Abstract 1391.
9. Jeyabalan A, Shelley-Fraser G, Medford AR. Impact of needle gauge on characterization of endobronchial ultrasound-guided transbronchial needle aspiration
(EBUS-TBNA) histology samples. Respirology. 2014;19:735–739.
10. Modi D, Copelen KL, Shilson E, et al.. Endobronchial ultrasound guided fine needle aspiration biopsies of the thoracic lymph nodes using 21 gauge or 22 gauge needles have similar cellular adequacies for diagnostic purpose - An observation. In: Proceedings of the American Thoracic Society International Conference; May 13-18, 2011; Denver, United States. Abstract A44.
11. Nakajima T, Yasufuku K, Takahashi R, et al.. Comparison of 21-gauge and 22-gauge aspiration needle during endobronchial ultrasound-guided transbronchial needle aspiration
. Respirology. 2011;16:90–94.
12. Oki M, Saka H, Kitagawa C, et al.. Randomized study of 21-gauge versus 22-gauge endobronchial ultrasound-guided transbronchial needle aspiration
needles for sampling histology specimens. J Bronchology Interv Pulmonol. 2011;18:306–310.
13. Saji J, Kurimoto N, Morita K, et al.. Comparison of 21-gauge and 22-gauge needles for endobronchial ultrasound-guided transbronchial needle aspiration
of mediastinal and hilar lymph nodes. J Bronchology Interv Pulmonol. 2011;18:239–246.
14. Moher D, Liberati A, Tetzlaff J, et al.. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097.
15. Jeyabalan A, Medford ARL. The effect of needle gauge on characterisation of histology samples at endobronchial ultrasound-guided transbronchial needle aspiration
(EBUS-TBNA). Thorax. 2012;67:A12–A13.
16. Jeyabalan A, Medford ARL. Adequacy of 22 and 21 Gauge EBUS-TBNA histology samples for genotyping of primary lung adenocarcinoma. Thorax. 2013;68:A199–A200.
17. Medford ARL, Bennett JA, Free CM, et al.. Endobronchial ultrasound guided transbronchial needle aspiration. Postgrad Med J. 2010;86:106–115.
18. Yasufuku K, Nakajima T, Motoori K, et al.. Comparison of endobronchial ultrasound, positron emission tomography, and CT for lymph node staging of lung cancer. Chest. 2006;130:710–718.
19. Herth F, Becker HD, Ernst A. Conventional vs endobronchial ultrasound-guided transbronchial needle aspiration
: a randomized trial. Chest. 2004;125:322–325.
20. Zhu T, Zhang X, Xu J, et al.. Endobronchial ultrasound guided-transbronchial needle aspiration vs. conventional transbronchial needle aspiration in the diagnosis of mediastinal masses: a meta-analysis. Mol Clin Oncol. 2014;2:151–155.