Chronic Respiratory Disease http://crd.sagepub.com/
Predictors of efficacy for endobronchial valves in bronchoscopic lung volume reduction: A meta-analysis Imran H Iftikhar, Franklin R McGuire and Ali I Musani Chronic Respiratory Disease 2014 11: 237 originally published online 21 August 2014 DOI: 10.1177/1479972314546766 The online version of this article can be found at: http://crd.sagepub.com/content/11/4/237
Published by: http://www.sagepublications.com
Additional services and information for Chronic Respiratory Disease can be found at: Email Alerts: http://crd.sagepub.com/cgi/alerts Subscriptions: http://crd.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://crd.sagepub.com/content/11/4/237.refs.html
>> Version of Record - Oct 31, 2014 OnlineFirst Version of Record - Aug 21, 2014 What is This?
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Original Article
Predictors of efficacy for endobronchial valves in bronchoscopic lung volume reduction: A meta-analysis
Chronic Respiratory Disease 2014, Vol. 11(4) 237–245 ª The Author(s) 2014 Reprints and permission: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1479972314546766 crd.sagepub.com
Imran H Iftikhar1, Franklin R McGuire1 and Ali I Musani2
Abstract Over the last several years, numerous trials have been carried out to check the efficacy of one-way valves in the management of advanced emphysema. While the design of the valves has not altered much, by selectively studying these valves in a select group of participants, such as those with and without intact fissures (FIþ and FI), and by using different procedural techniques, our understanding of the valves has evolved. In this metaanalysis, we sought to study the effect of these factors on the efficacy of one-way valves. From PubMed and Embase, we included only those studies that provided separate data on fissure integrity or collateral ventilation. Our study outcomes included the mean change in forced expiratory volume in first second (FEV1), 6-minute walk distance (6MWD) and the St George’s Respiratory Questionnaire (SGRQ). In the FIþ subgroup of participants, the pooled standardized mean difference in FEV1, 6MWD, and SGRQ were 0.50 (95% confidence interval (CI): 0.34 to 0.67), p 0.001, 0.29 (95% CI: 0.13 to 0.45), p 0.001 and 6.02 (95% CI: 12.12 to 0.06), p ¼ 0.05, respectively. In comparison, these results were superior to the FI subgroup of participants. A separate analysis of the FIþ subgroup based on lobar occlusion versus nonlobar occlusion favored the former for superior efficacy. The preliminary findings of our meta-analysis confirm that one-way valves perform better in a select group of patients who show intact fissures on lung imaging pretreatment and in those who achieve lobar occlusion. Keywords Endobronchial valves, emphysema, lung volume reduction, meta-analysis, chronic obstructive pulmonary disease, bronchoscopy
Introduction Although different pharmacological treatments exist for the general chronic obstructive pulmonary disease (COPD) patients, those with severe emphysema and poor lung function are often considered for additional surgical procedures such as the lung volume reduction (LVR) surgery (LVRS). Despite being the gold standard of the surgical alternatives for treatment of advanced emphysema, LVRS is associated with significant short-term morbidity and mortality as well as high costs.1,2 This has fuelled an accelerated interest in studying the less invasive bronchoscopic methods of LVR. Other methods studied include the bronchial thermal vapor ablation (BTVA) therapy and the airway bypass stents. While the former lacks longterm follow-up data, for the latter, evidence in the literature has essentially been negative.
The bronchoscopic placement of endobronchial valves allows air to leave but not enter emphysematous areas of the lung, resulting in their collapse. In patients with heterogeneous emphysema, this method allows the relatively healthier lung to function better. However,
1
Division of Pulmonary, Critical Care Department of Medicine, University Columbia, SC, USA 2 Division of Pulmonary, Critical Care Department of Medicine, National Jewish USA
and Sleep Medicine, of South Carolina, and Sleep Medicine, Health, Denver, CO,
Corresponding author: Imran H Iftikhar, Division of Pulmonary, Critical Care and Sleep Medicine, Department of Medicine, University of South Carolina, One Medical Park, Suite 300, Columbia, SC 29203, USA. Email: [email protected]
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
238
Chronic Respiratory Disease 11(4)
certain factors such as collateral ventilation (CV) and fissure integrity (FI), which denotes the presence or absence of complete fissures on lung imaging scans, have a bearing on the long-term success of procedures with one-way valves.3,4 Although the number of studies published on sealants and BTVA is fairly small in comparison with one-way valves, emerging evidence suggests that their effect is independent of factors such as FI or CV.5,6 The one-way valves have been studied for almost a decade; however, only few studies were specifically designed to address this issue.3,7 In this meta-analysis, we sought to systematically review all published studies on one-way valves and analyze the effect of FI on clinical outcomes.
Methods Search strategy and selection criteria We searched PubMed and Embase databases from their inception to December 6, 2013. We used combinations of the following keywords: ‘‘endobronchial valves’’, ‘‘one-way valves’’, ‘‘lung volume reduction’’ and ‘‘emphysema’’. Boolean operators (AND/OR) were used to pair key search words. The search from PubMed yielded all the studies included in this meta-analysis. To ensure a thorough search of the literature, we hand searched the reference lists of the included studies and previously published meta-analyses. For inclusion in our meta-analysis, we considered only those studies that reported data on the forced expiratory volume in one second (FEV1), 6-minute walk distance (6MWD), and the St George’s Respiratory Questionnaire (SGRQ). Prospective nonrandomized and randomized controlled trials (RCT) were included provided pre- and postintervention data or the mean difference (between pre- and postintervention) was available. We excluded consecutive case series, case reports, and retrospective studies. As a general rule, for multiple publications of the same trials, we intended to include only the most recent one. We also excluded studies that did not provide any separate data on FI or CV. One investigator (IHI) independently searched the studies and performed the final screening. There were no disagreements between investigators on the inclusion or exclusion of a study. Figure 1 summarizes the results of the selection process (details provided in supplementary material).
Data abstraction From the included studies, data were extracted for first author’s name, year of publication, number of
study participants, their age and sex distribution, presence or absence of FI (FIþ or FI), and/or CV. For the analysis, we recorded the mean of pre- and postintervention FEV1, 6MWD, and SGRQ with standard deviations (SDs) and where necessary the mean difference with SDs. For RCTs, comparing a valve with either control or an active comparator, we extracted data only for the cohort that received valves (details provided in supplementary material).
Quantitative data synthesis The mean changes (standardized mean difference or mean difference, where applicable) in the outcomes along with their 95% confidence intervals (CIs) were estimated by pooling the available data using comprehensive meta-analysis software version 2.2.064. Forest plots were constructed to analyze the results. Fixed effects methods were used to account for variance within the studies. Random effects methods were used to account for variance between and within the studies.8 Statistical heterogeneity was assessed with the I2 statistic.9 An I2 > 60% indicated significant heterogeneity. Where moderate to high heterogeneity was noticed, we reported the results in random effects model. To check for publication bias, we constructed funnel plots of effect size and standard error8,10 and also analyzed results using the Begg and Mazumdar’s rank correlation test and the Eggers test of the intercept.11 We chose to conduct this assessment on the effect of all studies (FIþ and FI) on FEV1, as all selected studies were included in this method.
Results A total of five studies qualified for inclusion in the meta-analysis.4,7,12–14 The total number of study participants analyzed in this meta-analysis was 267 (153 in FIþ group and 114 in FI group). Table 1 outlines the baseline characteristics of the study population. On average, study participants were >58 years old. The duration of postintervention follow-up included in our meta-analysis was between 1 and 12 months. There were a total of two RCTs. The pooled standardized mean difference in FEV1 was 0.50 (95% CI: 0.34 to 0.67), I2 0%, p 0.001 (Figure 2) and 0.18 (95% CI: 0.16 to 0.52), I2 68%, p ¼ 0.30 (Figure 3), for the subgroup of participants in FIþ and FI groups, respectively.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Iftikhar et al.
239
Total no, of citations searched: 37
Excluded after reading abstracts: 21
Total no, of articles shortlisted: 16
Excluded: No information on fissure integrity: 5 studies on Zephyr valves (3 with same cohorts) and 3 studies on IBV valves Case series: 3
Total no, of articles included in meta-analysis: 5
[a] Analysis of studies reporting no fissure integrity: 3
Comparison of clinical outcomes
[b] Analysis of studies reporting fissure integrity: 5 (2 in common with [A] 1 study excluded (no information on lobar occlusion)
Analysis of studies reporting lobar occlusion: 2
Comparison of clinical outcomes
Analysis of studies reporting non lobar occlusion: 2
Figure 1. Study selection process and trial flow.
The pooled standardized mean difference in 6MWD was 0.29 (95% CI: 0.13 to 0.45), I2 0%, p < 0.001 (Figure 4) and 0.21 (95% CI: 0.004 to 0.43), I2 24%, p ¼ 0.05 (Figure 5), for the subgroup of participants in FIþ and FI groups, respectively.
The pooled mean difference in SGRQ was 6.02 (95% CI: 12.12 to 0.06), I2 79%, p 0.05 (Figure 6) and 2.43 (95% CI: 6.18 to 1.32), I2 34%, p ¼ 0.20 (Figure 7), for the subgroup of participants in FIþ and FI groups, respectively.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
240
Chronic Respiratory Disease 11(4)
Table 1. Baseline characteristics of studies. Author Chung et al.
Duration of follow-up
Study design 12
Sciurba et al.4 Herth et al.14 Venuta et al.7 Herth et al.13
Prospective, single center, single cohort Multicenter RCT Multicenter RCT Prospective, non randomized, single center longitudinal study Multicenter non-RCT
Number of participants (% of males) a
Age Mean (SD) 72 (8)a
90 days
7 (57%)
12 monthsa 12 months 1, 3, and 5 years.
220 (60.4%)b 111 (68)b 33c
65.34 (6.83)b 59.7 (7.9) b 35–75
30 days
CV: 51 (51%) CVþ: 29 (45%)
CV: 63 (10) CVþ: 63 (9)
SD: standard deviation; RCT : randomized controlled trial; CV: group with collateral ventilation absent; CVþ: group with collateral ventilation present. a Only six participants completed study. b Data represent participants in intervention cohort. c At the end of first year, a total of 33 participants remain.
Study name
Chung 2010 Sciurba 2010 Herth 2012 Venuta 2012 Herth 2013
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
0.421 0.800 0.517 0.527 0.339 0.509
0.426 0.189 0.161 0.276 0.144 0.086
–0.414 0.430 0.203 –0.013 0.057 0.340
1.256 1.170 0.832 1.067 0.622 0.679
0.323 0.000 0.001 0.056 0.018 0.000 –2.00
–1.00
0.00
Decrease in FEV1
1.00
2.00
Increase in FEV1
Figure 2. Standardized mean difference in FEV1 in FIþ group. FEV1 indicates forced expiratory volume in first second. FIþ indicates study participants with intact fissures on CT scans prior to treatment with valves.
Study name
Herth 2012 Venuta 2012 Herth 2013
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
0.000 0.610 0.030 0.181
0.122 0.218 0.186 0.176
–0.239 0.183 –0.334 –0.164
0.239 1.037 0.394 0.527
1.000 0.005 0.871 0.303 –2.00
–1.00
0.00
Decrease in FEV1
1.00
2.00
Increase in FEV1
Figure 3. Standardized mean difference in FEV1 FI group. FEV1 indicates forced expiratory volume in first second. FI indicates study participants without intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Iftikhar et al.
241
Study name
Chung 2010 Sciurba 2010 Herth 2012 Venuta 2012 Herth 2013
Statistics for each study Point estimate
Standard error
Lower limit
0.281 0.182 0.371 0.632 0.228 0.294
0.416 0.166 0.156 0.283 0.142 0.083
–0.535 –0.143 0.066 0.078 –0.050 0.131
Point estimate and 95% CI
Upper limit
p-Value
1.097 0.507 0.677 1.186 0.506 0.456
0.500 0.273 0.017 0.025 0.108 0.000 –2.00
–1.00
0.00
Decrease in 6 MWD
1.00
2.00
Increase in 6 MWD
Figure 4. Standardized mean difference in 6MWD in the FIþ group. 6MWD indicates 6-minute walk distance. FIþ indicates study participants with intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Study name
Herth 2012 Venuta 2012 Herth 2013
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
0.167 0.504 0.063 0.214
0.123 0.212 0.186 0.111
–0.074 0.088 –0.301 –0.004
0.408 0.920 0.427 0.431
0.175 0.018 0.735 0.054 –2.00
–1.00
0.00
Decrease in 6 MWD
1.00
2.00
Increase in 6 MWD
Figure 5. Standardized mean difference in 6MWD in FI group. 6MWD indicates 6-minute walk distance. FI indicates study participants without intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Study name
Chung 2010 Sciurba 2010 Herth 2012 Herth 2013
Statistics for each study Point estimate
Standard error
Lower limit
–9.000 –5.400 0.000 –12.000 –6.028
10.638 2.045 2.261 2.223 3.109
–29.850 –9.408 –4.432 –16.357 –12.122
Upper limit 11.850 –1.392 4.432 –7.643 0.065
Point estimate and 95% CI p-Value 0.398 0.008 1.000 0.000 0.053 –30.00
–15.00
0.00
Decrease in SGRQ
15.00
30.00
Increase in SGRQ
Figure 6. Mean difference in SGRQ score in FIþ group. SGRQ indicates St George’s Respiratory Questionnaire. FIþ indicates study participants with intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
242
Chronic Respiratory Disease 11(4)
Study name
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
–1.000 –5.000
1.710 2.735
–4.352 –10.361
2.352 0.361
0.559 0.068
–2.430
1.917
–6.188
1.327
0.205
Herth 2012 Herth 2013
–12.00
–6.00 Decrease in SGRQ
0.00
6.00
12.00
Increase in SGRQ
Figure 7. Mean difference in SGRQ in FI group. SGRQ indicates St George’s Respiratory Questionnaire. FI indicates study participants without intact fissures on CT scans prior to treatment with valves.
Discussion
Funnel plot of standard error by difference in means Standard error
0.0 0.1 0.2 0.3 0.4 –2.0
–1.5
–1.0
–0.5
0.0
0.5
1.0
1.5
2.0
Difference in means
Figure 8. Funnel plot for assessment of publication bias.
Assessment of publication bias The funnel plot (Figure 8) did not show evidence of publication bias on visual inspection. Begg and Mazumdar’s rank correlation test showed Kendall’s t-b (corrected for ties) of 0.10000, with a one-tailed p value of 0.40. The Eggers test of the intercept showed the intercept (B0) at 0.09 (95% CIs: 1.59, 1.77), with a one-tailed p value of 0.43.
Sensitivity analysis The sensitivity analysis was carried out by selectively removing one study from the overall analysis to check for the influence of one particular study on overall results. In the six different analyses, all except one (the analysis of 6MWD in the subgroup of participants in FI group) did not reveal any outlier study. In the aforementioned subgroup analysis, the study by Venuta et al.7 was identified as the outlier. We also conducted separate analyses based on whether or not lobar occlusion was reported in studies including participants only in the FIþ group. Overall, the results favored lobar occlusion (Table 2).
To the best of our knowledge, this is the first metaanalysis that has systematically analyzed the effects of FI on the clinical outcomes of endobronchial valves. Although our meta-analysis was designed to study different types of one-way valves, all of the studies included in our meta-analysis happened to be those that studied endobronchial valves (in particular Zephyr1 valves). Our meta-analysis answers a focused clinical question, that is, does FI matter in the performance of endobronchial valves? However, as simple as it seems, the overall success of the endobronchial valves in clinical practice depends on much more than that. Factors such as the ability to isolate the target lobe, correct sizing of the valve, and correct deployment and density pattern of emphysema on lung imaging scans determine the effectiveness of the valves. A procedure that results in incomplete occlusion of the target lobe due to an inadequate seal can result in an unsuccessful procedure. In both the North American and European cohorts of Valve for Emphysema PalliatioN Trial (VENT), only 162 out of the 331 treated participants achieved lobar exclusion.3,4 While operator-dependent and procedure-related factors are difficult to predict and considered variable across the board, the pretreatment determination of certain other factors such as the presence or absence of FI can predict efficacy of the valves. FI on computerized tomography (CT) scans has also been used as a surrogate for determination of CV.6 Indeed, interlobar collateral ventilation is also present in the normal lungs; however, in the emphysematous lungs, this communication is extensive. In the normal lungs, this communication exists in the form of pores of Kohn15
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Iftikhar et al.
243
Table 2. Effect of lobar occlusion on clinical outcomes in patients with FI.
FEV1 6MWD SGRQ
Lobar occlusiona (SMD/MD with lower and upper CIs)
Nonlobar occlusiona (SMD with lower and upper CIs)
0.82 SMD (0.52 to 1.12), p 0.01 0.30 SMD (0.03 to 0.56), p ¼ 0.02 5.05 MD (8.53 to 1.57), p 0.01
0.25 SMD (0.03 to 0.55), p ¼ 0.08 0.01 SMD (0.30 to 0.27), p ¼ 0.93 0.37 MD (3.75 to 4.51), p ¼ 0.85
FI: fissure integrity; SMD: standardized mean difference; MD: mean difference in absolute units; CI: confidence intervals; FEV1: forced expiratory volume in first second; 6MWD: 6-minute walk distance; SGRQ: St Georges Respiratory Questionnaire. a Results are based on studies by Herth et al.14 and Sciurba et al.4
to accessory bronchioloalveolar communications16 and the communication ducts at the end of the respiratory bronchioles.17 As such, these channels of communication are considered high-resistant lowflow channels in the normal lungs. In the emphysematous lungs, different areas of heterogeneous ventilation are created that lead to different pressure gradients within the lungs. The recruitment of the dormant collateral channels (pores of Kohn and channels of Lambert and Martin) allow for some degree of pressure equalization in the emphysematous lobes of the lungs in order to maintain some functional gas exchange. The study participants in four out of the five studies included in our meta-analysis had FI assessed by high-resolution CT (HRCT) scans. In the most recent study by Herth et al., instead of using HRCT scans, the authors directly measured CV using the Chartis system (Pulmonx1). This system involves isolating the target lobe by inflating a balloon catheter in the opening of the target lobe in order to measure expiratory flow. Persistence of expiratory flow after balloon inflation suggests CV. The use of such standardized procedures to determine CV holds promise, since the assessment of FI based on CT scans can be variable. In a recent study by Koenigkam-Santos et al., the authors studied interobserver agreement between pulmonologists and general radiologists and that between experienced chest radiologists for the assessment of fissure completeness based on thin section multidetector CT scans (MDCT) for patients with severe emphysema who were being considered for endobronchial valve treatment.18 They found that the highest agreement was reached between experienced chest radiologists, whereas that between the pulmonologists and general radiologists was only fair to moderate.18 Although this evidence holds true for the visual analysis of CT scans, emerging data from recent studies indicate that automatic quantification of fissure completeness based on CT scans can be just as accurate and less time consuming and less tedious.19
Our meta-analysis reinforces the importance of complete occlusion for better efficacy of the endobronchial valves. In the FIþ subgroup of study participants (who had intact fissures prior to treatment with endobronchial valves), those who achieved complete lobar occlusion showed statistically significant improvement in outcomes compared with those who did not achieve lobar occlusion (Table 2). Although our meta-analysis did not include studies with intrabronchial valves (IBVs), this finding has also been noted in two studies on IBVs (Spiration1).20,21 In our meta-analysis, we were not able to study the effect of emphysema heterogeneity on the efficacy of endobronchial valves because different studies reported outcomes differently. In the VENT trial, the percent change in FEV1 between treatment and control arms was 5.9 (95% CIs:1.4 to 13.1), p 0.12 in the group with 15% heterogeneity. These results were not based on FI or lobar occlusion. Herth et al. studied the effect of heterogeneity on patients who showed FI and achieved lobar occlusion. In their analysis, 9 out of 20 participants had heterogeneity scores
Predictors of efficacy for endobronchial valves in bronchoscopic lung volume reduction: A meta-analysis Imran H Iftikhar, Franklin R McGuire and Ali I Musani Chronic Respiratory Disease 2014 11: 237 originally published online 21 August 2014 DOI: 10.1177/1479972314546766 The online version of this article can be found at: http://crd.sagepub.com/content/11/4/237
Published by: http://www.sagepublications.com
Additional services and information for Chronic Respiratory Disease can be found at: Email Alerts: http://crd.sagepub.com/cgi/alerts Subscriptions: http://crd.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://crd.sagepub.com/content/11/4/237.refs.html
>> Version of Record - Oct 31, 2014 OnlineFirst Version of Record - Aug 21, 2014 What is This?
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Original Article
Predictors of efficacy for endobronchial valves in bronchoscopic lung volume reduction: A meta-analysis
Chronic Respiratory Disease 2014, Vol. 11(4) 237–245 ª The Author(s) 2014 Reprints and permission: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1479972314546766 crd.sagepub.com
Imran H Iftikhar1, Franklin R McGuire1 and Ali I Musani2
Abstract Over the last several years, numerous trials have been carried out to check the efficacy of one-way valves in the management of advanced emphysema. While the design of the valves has not altered much, by selectively studying these valves in a select group of participants, such as those with and without intact fissures (FIþ and FI), and by using different procedural techniques, our understanding of the valves has evolved. In this metaanalysis, we sought to study the effect of these factors on the efficacy of one-way valves. From PubMed and Embase, we included only those studies that provided separate data on fissure integrity or collateral ventilation. Our study outcomes included the mean change in forced expiratory volume in first second (FEV1), 6-minute walk distance (6MWD) and the St George’s Respiratory Questionnaire (SGRQ). In the FIþ subgroup of participants, the pooled standardized mean difference in FEV1, 6MWD, and SGRQ were 0.50 (95% confidence interval (CI): 0.34 to 0.67), p 0.001, 0.29 (95% CI: 0.13 to 0.45), p 0.001 and 6.02 (95% CI: 12.12 to 0.06), p ¼ 0.05, respectively. In comparison, these results were superior to the FI subgroup of participants. A separate analysis of the FIþ subgroup based on lobar occlusion versus nonlobar occlusion favored the former for superior efficacy. The preliminary findings of our meta-analysis confirm that one-way valves perform better in a select group of patients who show intact fissures on lung imaging pretreatment and in those who achieve lobar occlusion. Keywords Endobronchial valves, emphysema, lung volume reduction, meta-analysis, chronic obstructive pulmonary disease, bronchoscopy
Introduction Although different pharmacological treatments exist for the general chronic obstructive pulmonary disease (COPD) patients, those with severe emphysema and poor lung function are often considered for additional surgical procedures such as the lung volume reduction (LVR) surgery (LVRS). Despite being the gold standard of the surgical alternatives for treatment of advanced emphysema, LVRS is associated with significant short-term morbidity and mortality as well as high costs.1,2 This has fuelled an accelerated interest in studying the less invasive bronchoscopic methods of LVR. Other methods studied include the bronchial thermal vapor ablation (BTVA) therapy and the airway bypass stents. While the former lacks longterm follow-up data, for the latter, evidence in the literature has essentially been negative.
The bronchoscopic placement of endobronchial valves allows air to leave but not enter emphysematous areas of the lung, resulting in their collapse. In patients with heterogeneous emphysema, this method allows the relatively healthier lung to function better. However,
1
Division of Pulmonary, Critical Care Department of Medicine, University Columbia, SC, USA 2 Division of Pulmonary, Critical Care Department of Medicine, National Jewish USA
and Sleep Medicine, of South Carolina, and Sleep Medicine, Health, Denver, CO,
Corresponding author: Imran H Iftikhar, Division of Pulmonary, Critical Care and Sleep Medicine, Department of Medicine, University of South Carolina, One Medical Park, Suite 300, Columbia, SC 29203, USA. Email: [email protected]
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
238
Chronic Respiratory Disease 11(4)
certain factors such as collateral ventilation (CV) and fissure integrity (FI), which denotes the presence or absence of complete fissures on lung imaging scans, have a bearing on the long-term success of procedures with one-way valves.3,4 Although the number of studies published on sealants and BTVA is fairly small in comparison with one-way valves, emerging evidence suggests that their effect is independent of factors such as FI or CV.5,6 The one-way valves have been studied for almost a decade; however, only few studies were specifically designed to address this issue.3,7 In this meta-analysis, we sought to systematically review all published studies on one-way valves and analyze the effect of FI on clinical outcomes.
Methods Search strategy and selection criteria We searched PubMed and Embase databases from their inception to December 6, 2013. We used combinations of the following keywords: ‘‘endobronchial valves’’, ‘‘one-way valves’’, ‘‘lung volume reduction’’ and ‘‘emphysema’’. Boolean operators (AND/OR) were used to pair key search words. The search from PubMed yielded all the studies included in this meta-analysis. To ensure a thorough search of the literature, we hand searched the reference lists of the included studies and previously published meta-analyses. For inclusion in our meta-analysis, we considered only those studies that reported data on the forced expiratory volume in one second (FEV1), 6-minute walk distance (6MWD), and the St George’s Respiratory Questionnaire (SGRQ). Prospective nonrandomized and randomized controlled trials (RCT) were included provided pre- and postintervention data or the mean difference (between pre- and postintervention) was available. We excluded consecutive case series, case reports, and retrospective studies. As a general rule, for multiple publications of the same trials, we intended to include only the most recent one. We also excluded studies that did not provide any separate data on FI or CV. One investigator (IHI) independently searched the studies and performed the final screening. There were no disagreements between investigators on the inclusion or exclusion of a study. Figure 1 summarizes the results of the selection process (details provided in supplementary material).
Data abstraction From the included studies, data were extracted for first author’s name, year of publication, number of
study participants, their age and sex distribution, presence or absence of FI (FIþ or FI), and/or CV. For the analysis, we recorded the mean of pre- and postintervention FEV1, 6MWD, and SGRQ with standard deviations (SDs) and where necessary the mean difference with SDs. For RCTs, comparing a valve with either control or an active comparator, we extracted data only for the cohort that received valves (details provided in supplementary material).
Quantitative data synthesis The mean changes (standardized mean difference or mean difference, where applicable) in the outcomes along with their 95% confidence intervals (CIs) were estimated by pooling the available data using comprehensive meta-analysis software version 2.2.064. Forest plots were constructed to analyze the results. Fixed effects methods were used to account for variance within the studies. Random effects methods were used to account for variance between and within the studies.8 Statistical heterogeneity was assessed with the I2 statistic.9 An I2 > 60% indicated significant heterogeneity. Where moderate to high heterogeneity was noticed, we reported the results in random effects model. To check for publication bias, we constructed funnel plots of effect size and standard error8,10 and also analyzed results using the Begg and Mazumdar’s rank correlation test and the Eggers test of the intercept.11 We chose to conduct this assessment on the effect of all studies (FIþ and FI) on FEV1, as all selected studies were included in this method.
Results A total of five studies qualified for inclusion in the meta-analysis.4,7,12–14 The total number of study participants analyzed in this meta-analysis was 267 (153 in FIþ group and 114 in FI group). Table 1 outlines the baseline characteristics of the study population. On average, study participants were >58 years old. The duration of postintervention follow-up included in our meta-analysis was between 1 and 12 months. There were a total of two RCTs. The pooled standardized mean difference in FEV1 was 0.50 (95% CI: 0.34 to 0.67), I2 0%, p 0.001 (Figure 2) and 0.18 (95% CI: 0.16 to 0.52), I2 68%, p ¼ 0.30 (Figure 3), for the subgroup of participants in FIþ and FI groups, respectively.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Iftikhar et al.
239
Total no, of citations searched: 37
Excluded after reading abstracts: 21
Total no, of articles shortlisted: 16
Excluded: No information on fissure integrity: 5 studies on Zephyr valves (3 with same cohorts) and 3 studies on IBV valves Case series: 3
Total no, of articles included in meta-analysis: 5
[a] Analysis of studies reporting no fissure integrity: 3
Comparison of clinical outcomes
[b] Analysis of studies reporting fissure integrity: 5 (2 in common with [A] 1 study excluded (no information on lobar occlusion)
Analysis of studies reporting lobar occlusion: 2
Comparison of clinical outcomes
Analysis of studies reporting non lobar occlusion: 2
Figure 1. Study selection process and trial flow.
The pooled standardized mean difference in 6MWD was 0.29 (95% CI: 0.13 to 0.45), I2 0%, p < 0.001 (Figure 4) and 0.21 (95% CI: 0.004 to 0.43), I2 24%, p ¼ 0.05 (Figure 5), for the subgroup of participants in FIþ and FI groups, respectively.
The pooled mean difference in SGRQ was 6.02 (95% CI: 12.12 to 0.06), I2 79%, p 0.05 (Figure 6) and 2.43 (95% CI: 6.18 to 1.32), I2 34%, p ¼ 0.20 (Figure 7), for the subgroup of participants in FIþ and FI groups, respectively.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
240
Chronic Respiratory Disease 11(4)
Table 1. Baseline characteristics of studies. Author Chung et al.
Duration of follow-up
Study design 12
Sciurba et al.4 Herth et al.14 Venuta et al.7 Herth et al.13
Prospective, single center, single cohort Multicenter RCT Multicenter RCT Prospective, non randomized, single center longitudinal study Multicenter non-RCT
Number of participants (% of males) a
Age Mean (SD) 72 (8)a
90 days
7 (57%)
12 monthsa 12 months 1, 3, and 5 years.
220 (60.4%)b 111 (68)b 33c
65.34 (6.83)b 59.7 (7.9) b 35–75
30 days
CV: 51 (51%) CVþ: 29 (45%)
CV: 63 (10) CVþ: 63 (9)
SD: standard deviation; RCT : randomized controlled trial; CV: group with collateral ventilation absent; CVþ: group with collateral ventilation present. a Only six participants completed study. b Data represent participants in intervention cohort. c At the end of first year, a total of 33 participants remain.
Study name
Chung 2010 Sciurba 2010 Herth 2012 Venuta 2012 Herth 2013
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
0.421 0.800 0.517 0.527 0.339 0.509
0.426 0.189 0.161 0.276 0.144 0.086
–0.414 0.430 0.203 –0.013 0.057 0.340
1.256 1.170 0.832 1.067 0.622 0.679
0.323 0.000 0.001 0.056 0.018 0.000 –2.00
–1.00
0.00
Decrease in FEV1
1.00
2.00
Increase in FEV1
Figure 2. Standardized mean difference in FEV1 in FIþ group. FEV1 indicates forced expiratory volume in first second. FIþ indicates study participants with intact fissures on CT scans prior to treatment with valves.
Study name
Herth 2012 Venuta 2012 Herth 2013
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
0.000 0.610 0.030 0.181
0.122 0.218 0.186 0.176
–0.239 0.183 –0.334 –0.164
0.239 1.037 0.394 0.527
1.000 0.005 0.871 0.303 –2.00
–1.00
0.00
Decrease in FEV1
1.00
2.00
Increase in FEV1
Figure 3. Standardized mean difference in FEV1 FI group. FEV1 indicates forced expiratory volume in first second. FI indicates study participants without intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Iftikhar et al.
241
Study name
Chung 2010 Sciurba 2010 Herth 2012 Venuta 2012 Herth 2013
Statistics for each study Point estimate
Standard error
Lower limit
0.281 0.182 0.371 0.632 0.228 0.294
0.416 0.166 0.156 0.283 0.142 0.083
–0.535 –0.143 0.066 0.078 –0.050 0.131
Point estimate and 95% CI
Upper limit
p-Value
1.097 0.507 0.677 1.186 0.506 0.456
0.500 0.273 0.017 0.025 0.108 0.000 –2.00
–1.00
0.00
Decrease in 6 MWD
1.00
2.00
Increase in 6 MWD
Figure 4. Standardized mean difference in 6MWD in the FIþ group. 6MWD indicates 6-minute walk distance. FIþ indicates study participants with intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Study name
Herth 2012 Venuta 2012 Herth 2013
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
0.167 0.504 0.063 0.214
0.123 0.212 0.186 0.111
–0.074 0.088 –0.301 –0.004
0.408 0.920 0.427 0.431
0.175 0.018 0.735 0.054 –2.00
–1.00
0.00
Decrease in 6 MWD
1.00
2.00
Increase in 6 MWD
Figure 5. Standardized mean difference in 6MWD in FI group. 6MWD indicates 6-minute walk distance. FI indicates study participants without intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Study name
Chung 2010 Sciurba 2010 Herth 2012 Herth 2013
Statistics for each study Point estimate
Standard error
Lower limit
–9.000 –5.400 0.000 –12.000 –6.028
10.638 2.045 2.261 2.223 3.109
–29.850 –9.408 –4.432 –16.357 –12.122
Upper limit 11.850 –1.392 4.432 –7.643 0.065
Point estimate and 95% CI p-Value 0.398 0.008 1.000 0.000 0.053 –30.00
–15.00
0.00
Decrease in SGRQ
15.00
30.00
Increase in SGRQ
Figure 6. Mean difference in SGRQ score in FIþ group. SGRQ indicates St George’s Respiratory Questionnaire. FIþ indicates study participants with intact fissures on CT scans prior to treatment with valves. FI: fissure integrity; CT: computed tomography.
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
242
Chronic Respiratory Disease 11(4)
Study name
Statistics for each study
Point estimate and 95% CI
Point estimate
Standard error
Lower limit
Upper limit
p-Value
–1.000 –5.000
1.710 2.735
–4.352 –10.361
2.352 0.361
0.559 0.068
–2.430
1.917
–6.188
1.327
0.205
Herth 2012 Herth 2013
–12.00
–6.00 Decrease in SGRQ
0.00
6.00
12.00
Increase in SGRQ
Figure 7. Mean difference in SGRQ in FI group. SGRQ indicates St George’s Respiratory Questionnaire. FI indicates study participants without intact fissures on CT scans prior to treatment with valves.
Discussion
Funnel plot of standard error by difference in means Standard error
0.0 0.1 0.2 0.3 0.4 –2.0
–1.5
–1.0
–0.5
0.0
0.5
1.0
1.5
2.0
Difference in means
Figure 8. Funnel plot for assessment of publication bias.
Assessment of publication bias The funnel plot (Figure 8) did not show evidence of publication bias on visual inspection. Begg and Mazumdar’s rank correlation test showed Kendall’s t-b (corrected for ties) of 0.10000, with a one-tailed p value of 0.40. The Eggers test of the intercept showed the intercept (B0) at 0.09 (95% CIs: 1.59, 1.77), with a one-tailed p value of 0.43.
Sensitivity analysis The sensitivity analysis was carried out by selectively removing one study from the overall analysis to check for the influence of one particular study on overall results. In the six different analyses, all except one (the analysis of 6MWD in the subgroup of participants in FI group) did not reveal any outlier study. In the aforementioned subgroup analysis, the study by Venuta et al.7 was identified as the outlier. We also conducted separate analyses based on whether or not lobar occlusion was reported in studies including participants only in the FIþ group. Overall, the results favored lobar occlusion (Table 2).
To the best of our knowledge, this is the first metaanalysis that has systematically analyzed the effects of FI on the clinical outcomes of endobronchial valves. Although our meta-analysis was designed to study different types of one-way valves, all of the studies included in our meta-analysis happened to be those that studied endobronchial valves (in particular Zephyr1 valves). Our meta-analysis answers a focused clinical question, that is, does FI matter in the performance of endobronchial valves? However, as simple as it seems, the overall success of the endobronchial valves in clinical practice depends on much more than that. Factors such as the ability to isolate the target lobe, correct sizing of the valve, and correct deployment and density pattern of emphysema on lung imaging scans determine the effectiveness of the valves. A procedure that results in incomplete occlusion of the target lobe due to an inadequate seal can result in an unsuccessful procedure. In both the North American and European cohorts of Valve for Emphysema PalliatioN Trial (VENT), only 162 out of the 331 treated participants achieved lobar exclusion.3,4 While operator-dependent and procedure-related factors are difficult to predict and considered variable across the board, the pretreatment determination of certain other factors such as the presence or absence of FI can predict efficacy of the valves. FI on computerized tomography (CT) scans has also been used as a surrogate for determination of CV.6 Indeed, interlobar collateral ventilation is also present in the normal lungs; however, in the emphysematous lungs, this communication is extensive. In the normal lungs, this communication exists in the form of pores of Kohn15
Downloaded from crd.sagepub.com at TEXAS SOUTHERN UNIVERSITY on November 20, 2014
Iftikhar et al.
243
Table 2. Effect of lobar occlusion on clinical outcomes in patients with FI.
FEV1 6MWD SGRQ
Lobar occlusiona (SMD/MD with lower and upper CIs)
Nonlobar occlusiona (SMD with lower and upper CIs)
0.82 SMD (0.52 to 1.12), p 0.01 0.30 SMD (0.03 to 0.56), p ¼ 0.02 5.05 MD (8.53 to 1.57), p 0.01
0.25 SMD (0.03 to 0.55), p ¼ 0.08 0.01 SMD (0.30 to 0.27), p ¼ 0.93 0.37 MD (3.75 to 4.51), p ¼ 0.85
FI: fissure integrity; SMD: standardized mean difference; MD: mean difference in absolute units; CI: confidence intervals; FEV1: forced expiratory volume in first second; 6MWD: 6-minute walk distance; SGRQ: St Georges Respiratory Questionnaire. a Results are based on studies by Herth et al.14 and Sciurba et al.4
to accessory bronchioloalveolar communications16 and the communication ducts at the end of the respiratory bronchioles.17 As such, these channels of communication are considered high-resistant lowflow channels in the normal lungs. In the emphysematous lungs, different areas of heterogeneous ventilation are created that lead to different pressure gradients within the lungs. The recruitment of the dormant collateral channels (pores of Kohn and channels of Lambert and Martin) allow for some degree of pressure equalization in the emphysematous lobes of the lungs in order to maintain some functional gas exchange. The study participants in four out of the five studies included in our meta-analysis had FI assessed by high-resolution CT (HRCT) scans. In the most recent study by Herth et al., instead of using HRCT scans, the authors directly measured CV using the Chartis system (Pulmonx1). This system involves isolating the target lobe by inflating a balloon catheter in the opening of the target lobe in order to measure expiratory flow. Persistence of expiratory flow after balloon inflation suggests CV. The use of such standardized procedures to determine CV holds promise, since the assessment of FI based on CT scans can be variable. In a recent study by Koenigkam-Santos et al., the authors studied interobserver agreement between pulmonologists and general radiologists and that between experienced chest radiologists for the assessment of fissure completeness based on thin section multidetector CT scans (MDCT) for patients with severe emphysema who were being considered for endobronchial valve treatment.18 They found that the highest agreement was reached between experienced chest radiologists, whereas that between the pulmonologists and general radiologists was only fair to moderate.18 Although this evidence holds true for the visual analysis of CT scans, emerging data from recent studies indicate that automatic quantification of fissure completeness based on CT scans can be just as accurate and less time consuming and less tedious.19
Our meta-analysis reinforces the importance of complete occlusion for better efficacy of the endobronchial valves. In the FIþ subgroup of study participants (who had intact fissures prior to treatment with endobronchial valves), those who achieved complete lobar occlusion showed statistically significant improvement in outcomes compared with those who did not achieve lobar occlusion (Table 2). Although our meta-analysis did not include studies with intrabronchial valves (IBVs), this finding has also been noted in two studies on IBVs (Spiration1).20,21 In our meta-analysis, we were not able to study the effect of emphysema heterogeneity on the efficacy of endobronchial valves because different studies reported outcomes differently. In the VENT trial, the percent change in FEV1 between treatment and control arms was 5.9 (95% CIs:1.4 to 13.1), p 0.12 in the group with 15% heterogeneity. These results were not based on FI or lobar occlusion. Herth et al. studied the effect of heterogeneity on patients who showed FI and achieved lobar occlusion. In their analysis, 9 out of 20 participants had heterogeneity scores
