The Effects of Adjunctive Thoracic Epidural Blockade Therapy in Patients with Heart Failure : A Systematic Review and Cumulative Meta-Analysis of Open-Label Trials

Background: Accumulative evidences have revealed the cardiac protective effects of thoracic epidural blockade (TEB), an anesthetic and analgesic method widely used in medical practice. However, TEB doesn't receive worldwide consensus in heart failure (HF) treatment. Thus, we conducted the current systematic review and cumulative meta-analysis to evaluate the therapeutic effects of adjunctive TEB in patients with HF.Methods: We conducted searches of the published literature in MEDLINE, PsycINFO, Scopus, EMBASE, CENTRAL, CNKI, VIP, WanFang databases from inception to October 2014. Clinical trials that studied adjunctive TEB in HF treatment were analyzed. Effect sizes were summarized using odds ratio (OR) or weighted mean differences (WMDs) by suitable effect model, and secondary findings in each trial were comprehensively reviewed.Results: Eighteen eligible trials with 1076 subjects were identified. Benefits of adjunctive TEB in HF treatment were shown on improving New York Heart Association (NYHA) class (OR, 11.39; 95% confidence interval [CI], 6.41 to 20.24, P<0.00001), left ventricular ejection fraction (LVEF, WMD, 8.45%; 95% CI, 6.48 to 10.41, P<0.00001) and left ventricular end-diastolic diameters (LVEDd, WMD, -4.39 mm; 95% CI, -5.95 to -2.83, P<0.00001). The secondary findings of cardiac function and structure improvement in all trials were reported dominantly advantageous. No TEB-related serious adverse events were reported in these included trials.Conclusions: Present data support that adjunctive TEB might be a promising candidate for HF. However, the pooled results might be weakened by the study design, suggesting that further randomized controlled trials are needed.  Citation: Huang-Hui Wu, Hai-Yan Wang, Xiao-Hui Tan, Da-Xiu Wei, Hui-Hua Ke, Yan Zhang, et al. The effects of adjunctive thoracic epidural blockade therapy in patients with heart failure: a systematic review and cumulative meta-analysis of open-label trials. J Anesth Perioper Med 2015; 2: 301-15. doi: 10.24015/JAPM.2015.0041This is an open-access article, published by Evidence Based Communications (EBC). This work is licensed under the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium or format for any lawful purpose. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

creased plasma levels of norepinephrine (NE), angiotensin II (Ang II), and endothelin-1 (ET-1) (3).Initially, the enhanced activity of these neurohormonal systems facilitates to compensate the reduced blood pressure and cardiac output.However, persistent increasing workload of the heart and maladaptive cardiac remodeling might contribute to the cardiovascular death after a long term exposure to high levels of circulating catecholamines (CAs) and Ang.All these pathologic changes contribute to the worsened myocardial function, as well as accelerating left ventricular (LV) remodeling and promoting a vicious circle (4)(5)(6)(7).Considering the imbalance activity between sympathetic and parasympathetic nervous systems plays a crucial role on the opposing influences on cardiac function, thus, to address the imbalance is essential in finding effective new strategies of intervention (8,9).
Thoracic epidural blockade (TEB) is a worldwide anesthetic and analgesic method.Growing experimental and clinical evidences have confirmed its cardiac protective effect on improving LV function in coronary artery disease, unstable angina pectoris, as well as myocardial infarction (MI) (10)(11)(12)(13)(14)(15).Theoretically, it is possible that the blockade of hyper-activated cardiac sympathetic nervous systems by TEB could reduce the harmful effects of excessive and persistent imbalance between cardiac sympathetic and parasympathetic nervous systems, as well as rebalancing the cardiac oxygen metabolism due to decreased heart rate (HR) and afterload (16), and then, benefiting for the patients with HF (17).However, the adverse events related to TEB in medical practice are as important as that of its benefits.Due to its invasive nature, use of a TEB technique may be accompanied by known complications, such as trauma, bleeding, hematoma, infection, thrombosis, embolism, spinal injury and ligament injury.Considering the clinical pros and cons, the current TEB therapy doesn't receive worldwide consensus in HF treatment and the specific underlying therapeutic mechanisms still remain unclear.Thus, we conducted the current systematic review and cumulative meta-analysis to identify the effects of adjunctive TEB therapy in patients with HF.

METHODS
The current systematic review and meta-analysis was performed based on the Quality of Reporting of Meta-analyses (QUORUM) guidelines (18) and the recommendations of the Cochrane Collaboration (19).

Data Source and Search Strategy
The foreign electronic databases screened were MEDLINE (1970  Search was limited to human and performed for all languages, year and type of publication.The search terms were ("[high] thoracic epidural [sympathetic] blockade" OR "[high] thoracic epidural anesthesia" OR "cardiac sympathetic blockade") AND ("heart failure" OR "cardiac insufficiency" OR "cardiac dysfunction") (APPENDIX).An independent hand search in reference sections of included trials, published meta-analyses, and relevant review articles was further conducted to identify additional publications.All the searches were performed by two of us (Huang-Hui Wu and Xiao-Hui Tan) independently.

Studies Selection
A study was eligible for inclusion if the following criteria were met: 1) Any randomized controlled trial (RCT), controlled clinical trial, or open label trial (OLT) designed with at least two groups that one control group receiving routine anti-HF therapy, and the other group receiving TEB with or without routine anti-HF therapy in patients with HF caused by any kind of disease; 2) TEB administrated with any type of strategy and local anesthetics; 3) Trials revealed at least one of primary or secondary outcomes men- If the data were unavailable in the article, the corresponding authors were contacted for missing information.If the outcomes in the pub lished studies were presented in a graph manner without any description of absolute value, Im -age J software Version 2.1.4.7 (Image J software, National Institutes of Health, Bethesda, MD, USA, http://imagej.nih.gov) was used to restore the missing data if we could not get the original data from the authors.
All data were independently extracted by 2 reviewers (Huang-Hui Wu and Xiao-Hui Tan), and then were checked and entered into a standard data collection form by the other 2 reviewers (Da-Xiu Wei and Hai-Yan Wang).All discrepancies were rechecked and consensus was reached by discussion with a third author (Hui-Hua Ke) involved.

Assessment of Quality and Risk of Bias
A critical evaluation of the included studies quality was performed by 2 reviewers (Da-Xiu Wei and Hai-Yan Wang) by using a 5-point Jadad scale (20).The main categories consisted of the following 5 items: "Was the study described as randomized?(1 or 0)", "Was the method used to generate the sequence of randomization described and appropriate (random numbers, computer-generated, etc.)? (1 or 0)", "Was the study described as double-blind?(1 or 0)", "Was the method of double-blinding described and appropriate (identical placebo, active placebo, dummy, etc.)? (1 or 0)", and "Was there a description of withdrawals and drop-outs?(1 or 0)".A score of 4 to 5 was considered a high methodological quality.Cohen's kappa was applied for calculating inter-rater agreement.
Two reviewers (Hui-Hua Ke and Yan Zhang) independently evaluated the risk of bias according to the recommendations from the Cochrane collaboration.The main categories consisted of selection bias, performance bias, detection bias, attrition bias, reporting bias, and other bias.Each domain was assessed to "high risk", "low risk", or "unclear".

Assessment of Heterogeneity and Publication Bias
The NYHA class, LVEF, and LVEDd were pooled to calculate the effect size.And then the study heterogeneity at overall level was investigated by using a χ 2 test and calculating I 2 statistic (21).When I 2 was 50% or lower, a low heterogeneity was rated and the data were pooled with a fixed effect model.When I 2 was large than 50% , a significant heterogeneity was rated and the da- ta were pooled with a random effects model (21).Subgroup analysis was further used to identify the significant heterogeneity according to study quality (Jadad score <3 and 3 or more), number of patients (<60 and 60 or more), age (<50 years and >50 years or more), course of therapy (<2 weeks and 2 weeks or more), frequency (<6/day and 6/day or more), category of HF (ischemic, congenital, idiopathic, and 2 or more), and therapy (TEB only and TEB + Routine).Furthermore, meta-regression was used to identify the origin of heterogeneity, such as the different age, course, frequency, category, and therapy.
We performed the sensitivity analyses to examine the effect of LVEF and LVEDd by excluding studies with low quality or high risk of bias, and investigated the potential publication bias by using graphical (Begg's funnel plot) (22) and statistical tests (Egger's test) (23).

Statistical Analysis
Binary variables (NYHA class) were pooled by using odds ratio (OR) with 95% confidence intervals (CIs).If the 95% CI covered the value of 1, we considered that the difference between TEB and control group was not statistically significant.A L'Abbé plot was built by plotting the observed NYHA class improvement rates of the TEB group against the placebo group.If adjunctive TEB is more effective in improving NYHA class, the improvement rates of the TEB group will be higher than the placebo group.However, if the studies are distributed along the line of equality, significant interactions between effectiveness of adjunctive TEB and NYHA class improvement are unlikely (24,25).Continuous variables (LVEF and LVEDd) were pooled by using weighted mean difference (WMD) with their 95% CIs because they were measured by the same scale.If the 95% CI covered the value of 0, we considered that the difference between TEB and control group was not statistically significant.If these continuous data were only reported as mean or median with standard error or range, we converted them into mean with standard deviation (SD) as previously reported (26).The other secondary findings were systematically reviewed and presented as mean difference with P values considering the heterogene-ity.In the cumulative meta-analysis, outcome data for NYHA class, LVEF, and LVEDd from all available studies were analyzed sequentially according to the publication year in which they first became available.Statistical analyses were performed using STATA 12.0 software (Stata Corporation, College Station, TX, USA).

Methodological Quality and Risk of Bias
The Jadad score of each included study was presented in supplementary table 1.The median quality score was 3 (range from 2 to 4).Inter-rater reliability for this assessment was κ=0.73.
The risk of bias of included studies was presented in supplementary table 2. In total, 3 trials (17% ) clearly described the procedure of randomization, and none of the trials reporting the methods of allocation concealment.All trials were open-labeled for participants and personnel, and influenced by the detection bias.One trial (5% ) was rated with high risk of attribution bias.All trials had an unclear risk of bias on selective outcome reporting.And one trial (5% ) had high risk on other sources of bias.

Effects of Adjunctive TEB on Primary Outcomes
Data from 276 subjects was used to assess the therapeutic effect of adjunctive TEB based on NYHA class improvement.The meta-analysis revealed that adjunctive TEB was associated with a significant improvement of NYHA class compared with placebo group (Figure 1A; OR, 11.39; 95% CI, 6.41 to 20.24, P<0.00001).The I 2 value of 28.4% indicated without evidence of heterogeneity.On cumulative meta-analysis, the original significant result for an adjunctive TEB effect persisted with smaller 95% CI (Figure 1B).Similarly, the L'Abbé plot showed that the distribution of the trials was not parallel to the line of equality, suggesting that interactions between effectiveness of adjunctive TEB and NYHA class improvement were apparent (Figure 2).Basically, the rate of TEB successfully improved HF patients' NYHA class was 96.5% in a total of 312 participants vs. 66.8% in 298 patients receiving routine anti-HF therapy.
In a further subgroup analysis, compared with 19.2% of cases in a total of 203 patients receiving anti-HF routine therapy, 54.2% of cases in a total of 217 patients receiving adjunctive TEB improved their NYHA class more than 2 degree (OR, 8.78; 95% CI, 3.56 to 21.65, P= 0.01), while compared with 36.0% of cases in a total of 203 patients receiving anti-HF routine therapy, 41.7% of cases in a total of 217 patients receiving adjunctive TEB improved their NYHA class more than 1 but less than 2 degree (Figure 3; OR, 0.84; 95% CI, 0.44 to 1.62, P= 0.61).These findings successfully indicated the superiority of adjunctive TEB in symptom alleviation and quality of life improvement in patients with HF during a 1-4 weeks treatment.

Effects of Adjunctive TEB on Secondary Findings
Data from 830 subjects was used to assess the effect of adjunctive TEB on LVEF.The meta-analy-  Systematic Review and Meta-Analysis sis revealed that adjunctive TEB was associated with a significant increase of LVEF compared with placebo group (Figure 4A; WMD, 8.45% ; 95% CI, 6.48 to 10.41, P<0.00001).The I 2 value of 73.2% indicated significant heterogeneity.On cumulative meta-analysis, the original significant result for an adjunctive TEB effect persisted with smaller 95% CI (Figure 4B).Data from 683 subjects was used to assess the effect of adjunctive TEB on LVEDd.The metaanalysis revealed that adjunctive TEB was associated with a significant decrease of LVEDd compared with placebo group (Figure 5A; WMD, -4.39 mm; 95% CI, -5.95 to -2.83, P<0.00001).The I 2 value of 65.6% indicated significant heterogeneity.On cumulative meta-analysis, the original significant result for an adjunctive TEB effect persisted with smaller 95% CI (Figure 5B).
Data from 1002 subjects was used to assess the effect of adjunctive TEB on cardiac function.Cardiac functions including SV, E wave, A wave, E/A, LVFS, LVET, CO, CI, and NYHA were well summarized in supplementary table 3. We didn't perform a meta-analysis on them because of great heterogeneity.All studies provided similar evidences that TEB facilitated to improve HF patients' cardiac function evaluated by echocardiography compared with control group.
Data from 525 subjects was used to assess the effect of adjunctive TEB on cardiac structure.Cardiac structures including LVESd, LAEDs, RAEDs, RVEDd, LVEDV, LVESV, LVWI, and SI were well summarized in supplementary table 4. As the same, we didn't perform a meta-analysis   Systematic Review and Meta-Analysis (28) report that the all-causes mortality rate of patients in the TEB group was 21.5% , and that of the control group was 34.7% (hazard ratio: 0.548, P<0.01), and the rehospitalization rate in the TEB group and control group was 34% and 61% , respectively, in a follow-up of 18 ± 16 months.None of included trials reported severe TEB-related adverse events.Five trials (27,30,33,35,45) 6B).No other significant differences were identified in the predefined factors between TEB and control group both in LVEF and LVEDd (Figure 6A and 6B).
Results from sensitivity analysis revealed that dominant heterogeneity rated in Zhao et al. 2007 (34), Wang et al. 2008 (37), Wang et al. 2009 (38) in LVEF, and Ji et al. 2008 (36), Wang et al. 2008 (37), and Wang et al. 2009 (38) in LVEDd, we removed these from the analysis to produce statistically homogeneous estimates.In all cases, the homogeneous estimate did not differ materially in size from the original estimate.However, after sequential exclusion of each trial from all pooled analysis, the pooled effect size was not affected by the exclusion of any specific trial both in LVEF and LVEDd (Supple- Adjunctive Thoracic Epidural Blockade for Heart Failure Huang-Hui Wu et al. mentary Figure 3).Results from meta-regression revealed that neither age (Padjusted=0.99for LVEF and 0.60 for LVEDd), course (Padjusted=0.62for LVEF and 0.97 for LVEDd), frequency (Padjusted=0.56for LVEF and Padjusted=1.00for LVEDd), category (Padjusted=0.97for LVEF and Padjusted=1.00for LVEDd), nor therapy (Padjusted=1.00for LVEF and 1.00 for LVEDd) was contributed to the study heterogeneity.

Publication Bias
No potential publication bias was found in NYHA class, LVEF, and LVEDd according to both Begg's funnel plot (Supplementary Figure 4) and Egger's test (Supplementary Table 5).

Summary of Current Evidences
Our systematic review aimed to assess the therapeutic effects of adjunctive TEB on patients with HF.According to our screen strategy and inclusion criteria, 1076 active participants from 18 studies were reviewed.Generally, although our pooled results were extremely restricted by the risk of bias and quality of included studies, adjunctive TEB facilitated to improve NYHA class, enhance LVEF and reduce LVEDd in patients with HF.And TEB technique was well-tolerated and the rate of TEB-related adverse events was low.The major issues reported in these thials were headache and dizziness, swelling on puncture site, tiredness, cough, hypotension, and numbness on upper limbs.However, all these adverse events were mild to moderate and easily reversed.
Thus, we highlight that adjunctive TEB could be a safe and effective technique involved in the anti-HF therapy strategy.These superiorities derive from the specific therapeutic effects as following: i) TEB effectively interrupts the positive feedback loop between the increasing myocardial oxygen demand and decreasing flow of coronary blood induced by high sympathetic activation, and then, rebalances the cardiac oxygen metabolism (46,47).Nygård, et al. (48) found that TEB intervention could significantly reverse the cold pressor test-induced increase of coronary vascular resistance both in nonstenotic and stenotic coronary vessels, suggesting that TEB could, at least partly, normalize the myocardial blood flow response to sympathetic stimulation.ii) TEB effectively slows the maladaptive cardiac remodeling process.During HF, the ratio of beta adrenergic receptor (β-AR) expression is shifting.The prominent role of β2AR signaling and selective down regulating β1AR expression would contribute to the disease condition and cardiac remodeling (49,50).However, surprisingly, TEB intervention could partly reverse the ratio of β1/β2AR (49,50).In addition to the change of β1/β2AR-mediated signaling in HF process, our previous pre-clinical study also revealed the underlying mechanism of TEB ameliorating ventricular remodeling might involve the decreases of β3AR expression in a HF rat model induced by MI (51).iii) TEB presents an inhibitory effect on platelet aggregation and antithrombotic effects (52,53).These benefits possibly derive from the protection of endothelial cells (54)(55)(56).iv) TEB also increases the endocrinal to pericardial blood flow ratio and reduces the severity of myocardial ischemia, the stress response and pain experience, and the incidence of malignant ventricular arrhythmias (57)(58)(59).

Strengths and Limitations
The current systematic review provides some practical experiences of TEB in HF treatment, and we also made our best effort to minimize the potential bias by conducting an extensive search through multiple databases and successfully receiving part of original data from several authors.However, some potential limitations should be addressed.
On one hand, some limitations derive from the internal and external validity of the included trials.Firstly, numerous disadvantages exist with the open-labeled design in all included published trials to assess the efficiency of TEB in HF practice, although RCTs are considered as the gold standard in the medical literature.Notorious limitations are as following: inadequate sample size, short period of follow-up to detect the long-term prognosis, and difficulties in maintaining blinding and allocation concealment.Secondly, great clinical heterogeneity was rated in baseline demographics, category of HF, as well as course of TEB.Thus, we further performed subgroup analysis, sensitivity analysis, and meta- Systematic Review and Meta-Analysis regression to identify the heterogeneity.Thirdly, poor reporting main exclusion criteria in several trials restricted the generalizability of findings.Finally, since nearly all clinical investigations were performed in Asia, especially in China, the enrolled participants were predominantly Chinese, thus, the correlation analysis between identification of other demographic and clinical characteristic are needed to better identify a clinical "niche" in research outside of China.
On the other hand, some limitations are related to the meta-analytical approach.One is that some data were given only in graph form instead of original one.Image J was used to restore the related data when we failed to contact the correspondence authors.Thus, these possible sources of heterogeneity could not be examined.The other one is that we used I 2 for assessing the amount of heterogeneity in random-effects meta-analysis (60) and no statistically significant heterogeneity revealed, however, this method is very specific but not too sensitive.

Direction for Further Research
Our cumulative meta-analysis provides the currently limited data of the significant improve-ments of NYHA class, LVEF and LVEDd after TEB treatment, however, these evidences might be weakened by those limitations summarized above.Therefore, further researches using valid and reliable assessments are needed to more comprehensively evaluate the treatment effects of TEB on HF.Besides, more in-depth studies that compare different aspects of TEB and longterm outcome are needed, as well.
Further RCTs should be rigorously designed with special attention paid to blinding of assessors and sample sizes that are adequately powered.Furthermore, comprehensive patient characteristics and TEB strategies are required among different populations.And comparisons of TEB with routine anti-HF therapy with cost-effectiveness analysis would also facilitate to establish this treatment as a promising adjuvant option.

Cumulative meta-analysis of LVEF Figure
4. Forest Plot: Meta-Analysis (A) and Cumulative Analysis (B) for LVEF.