Bifurcation percutaneous coronary intervention (PCI) remains technically challenging and is considered high risk to procedural complications, particularly side branch (SB) occlusion. The jailed wire technique (JWT) is commonly recommended for SB protection; however, it does not fully prevent the complication. The jailed balloon technique (JBT) was introduced to provide more active protection, but previous evidence has been inconsistent. This study aimed to provide updated evidence on the effectiveness of JBT versus JWT for SB protection during bifurcation PCI. PubMed, ScienceDirect, and the Cochrane Library were searched for randomized trials and cohort studies comparing JBT and JWT in patients undergoing bifurcation PCI. The primary end points were procedure-related adverse events, including SB occlusion, coronary dissection, and peri-procedural myocardial infarction. Secondary end points included all-cause mortality, cardiac death, target-lesion revascularization, and major adverse cardiovascular events (MACE). Odds ratios with 95% confidence intervals were calculated as a measure. Study quality was assessed using the Newcastle–Ottawa Scale and risk of bias 2.0 tool. Ten studies involving 2,329 patients were included (793 JBT and 1,536 JWT). The risk of SB occlusion was lower in JBT group compared to JWT group (odds ratio = 0.31, 95% confidence interval 0.20 to 0.48; p <0.00001). There were no significant differences between the 2 techniques in coronary dissection, peri-procedural myocardial infarction, all-cause mortality, cardiac death, target-lesion revascularization, or MACE. All studies were considered good quality. In conclusion, the use of JBT was associated with a reduced risk of SB occlusion during bifurcation PCI, while both techniques demonstrated comparable outcomes for other procedural and long-term clinical outcomes.
Coronary bifurcation percutaneous coronary intervention (PCI) remains a challenging procedure and accounts for approximately 15% to 20% of all PCI cases. This procedure is known associated with intra and postprocedural complications. Stent deployment in the main vessel (MV) can result in plaque redistribution or carina shift, potentially compromising the side branch (SB). These changes may lead to complications, ranging from SB occlusion and peri-procedural myocardial infarction (MI) to adverse long-term clinical events. To reduce these risks, the European Bifurcation Club recommends the jailed wire technique (JWT) as a strategy for SB protection during bifurcation PCI. This technique involves leaving a guidewire in the SB while a stent is deployed in the MV. The jailed wire helps maintain SB patency and serves as a landmark for SB access in the event of occlusion. However, JWT does not provide complete SB protection, as the limited space occupied by the wire cannot prevent plaque migration or protrusion into the SB. To address these limitations, Burzotta et al introduced the jailed balloon technique (JBT) to enhance SB protection during bifurcation PCI. The JBT offers several advantages over the JWT. The jailed balloon can be inflated within the SB when necessary. Moreover, the presence of an uninflated jailed balloon may help reduce mechanical plaque migration toward the SB during MV stent deployment. These theories were confirmed by Jeremiah et al, which demonstrated that the use of JBT in coronary bifurcation lesions was associated with lower rates of SB loss and major adverse cardiac events (MACE). Nevertheless, previous studies comparing JBT and JWT in bifurcation PCI have yielded inconsistent findings. Although a previous meta-analysis suggested that JBT was more effective than JWT in preventing SB occlusion, it was limited by a small sample size and did not include the most recent evidence. Therefore, this systematic review and meta-analysis aim to provide updated evidence on the effectiveness of JBT compared with JWT for SB protection during bifurcation PCI.
Methods
We conducted this study based on an updated version of the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocol 2020.
Eligibility criteria
The inclusion criteria comprised studies that met all of the following conditions: (1) randomized trial or cohort, either prospective or retrospective in design, (2) patient with coronary bifurcation lesion underwent PCI, and (3) using JBT or JWT for SB protection. JBT is performed either with conventional JBT or modified/active JBT. Balloon-stent kissing technique (BSKT) is performed after JBT whenever necessary. JWT is performed by placing coronary guidewire in both MV and SB. MV was predilated, then MV stent was placed and dilated, while the SB guidewire was left in the vessel as protection. The primary end point of this analysis was procedure-related adverse events, including SB occlusion, coronary dissection, and peri-procedural MI. The secondary end points were long-term adverse events such as all-cause mortality, cardiac death, target-lesion revascularization (TLR), and MACE. Studies were excluded if they were duplicates, case reports or case series, review articles, abstracts, commentaries, editorials, low-quality studies, using techniques other than JBT or JWT, or studies with no outcome of interest.
Data sources and search strategy
We conducted a data search up until December 1, 2025 in PubMed, ScienceDirect, and Cochrane Library using keywords “side branch protection OR jailed wire technique OR jailed balloon technique OR coronary artery bifurcation.” Our search strategy did not mention a specific outcome or end point to broaden the search area, since this field has not been widely studied. None of the filters were applied in the search on the databases. We straighten out any search issues with a discussion.
Data extraction and bias assessment
Three independent reviewers screened the titles and abstracts of studies relevant to the research question. The screening results were then manually assessed by 2 additional reviewers based on the predetermined inclusion and exclusion criteria. For all eligible studies, full-text data were extracted and compiled into a Google Sheet. The extracted information included the first author’s name, year of publication, study design, population characteristics, reported end points, bifurcation location, admission diagnosis, follow-up duration, and study location. Following data compilation, relevant data aligned with the defined end point were extracted and entered into Review Manager software for further analysis. The quality of each included study was assessed using the Newcastle–Ottawa Scale (NOS) for cohort and risk of bias (RoB) 2.0 for randomized trials. Only studies with good quality by NOS or low RoB by RoB 2.0 were included in the analysis.
Statistical analysis
The extracted data were analyzed using Cochrane Collaboration Review Manager version 5.4. We assessed each parameter with odds ratios (ORs) using the Mantel–Haenszel method. All effect estimates were reported with 95% confidence intervals (CIs), and the results were visualized using forest plots. Statistical significance was defined as a p-value <0.05, with significance also considered when the 95% CI did not cross the middle line on the forest plot. Heterogeneity across studies was assessed using the I² statistic and Cochrane’s Q test. An I² value of <50% indicated low or no significant heterogeneity, and a fixed-effect model was applied. In cases when I² > 50%, indicating substantial heterogeneity, a random-effects model was used. A p-value of <0.05 for Cochrane’s Q test was considered statistically significant for heterogeneity. For outcomes with high heterogeneity (I² > 50%), a sensitivity analysis was conducted using the leave-one-out method, in which each study was excluded sequentially to assess the impact on overall heterogeneity and to identify potential sources of variability. Due to our inclusion criteria, subgroup analysis was conducted for randomized trials included in our studies. Funnel plots were generated for all end points to visually assess potential publication bias based on the symmetry of study distributions. ,
Results
The literature search yielded 1,798 records from 3 databases. After screening the title and/or abstract, 22 articles were assessed for eligibility. Articles with no full text available, non-JBT or JWT PCI, and that did not include our outcome of interest were excluded. Finally, 10 studies were considered eligible and underwent meta-analysis ( Figure 1 ).
Flow diagram. JBT = jailed balloon technique; JWT = jailed wire technique; PCI = percutaneous coronary intervention.
Patient characteristics
As seen in Table 1 , 10 included studies consisted of 4 randomized trials and 6 retrospective cohort studies. The total 2329 patients with coronary bifurcation lesions were grouped into JBT group (n = 793) and JWT group (n = 1536), which was dominated by male gender. In the JBT group, 375 (47.2%) patients underwent modified JBT as an SB protection, and 78 (9.8%) patients underwent further BSKT. From 10 included studies, 2 of them only included nonleft main bifurcation lesions. The most common Medina classification for bifurcation lesions across studies was 1.1.1 (53.6%). Six included studies specified the stent type, mainly drug-eluting stents (everolimus, zotarolimus, rapamycin, or paclitaxel). The shortest follow-up period was 1 month by Dou et al and Lai et al, while 4 of included studies did not mention their specific follow-up duration. The clinical follow-up data were collected by outpatient visit or by telephone call.
Table 1
Study characteristics
| Study | Design | Population (n) | Endpoint | Age (JBT/JWT) | Diagnosis (JBT%/JWT%) | Bifurcation (JBT%/JWT%) | Comorbidities (JBT%/JWT%) | JBT-group Vessel Size (mm) | JWT-group Vessel Size (mm) | Follow-up period | Location |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Lai et al | Retrospective cohort | JBT (17) & JWT (17) | MI, stent thrombosis, recurrent angina, all-cause mortality | Not mentioned | Not mentioned | Not mentioned | Not mentioned | Not mentioned | Not mentioned | 1 & 6 month | China |
| Kuno et al | Retrospective cohort | JBT (44) & JWT (506) | SB occlusion, MI, all-cause mortality, coronary dissection, perforation | 69.1 ± 8.3/67.8 ± 9.7 | STEMI (0%/20%), NSTEACS (22.5%/23%), Elective (75%/57.7%) | LM (11.3%/32.6%), LAD-D1 (79.5%/48.2%), LCX-OM (4.5%/9.7%), RCA-PDA (2.3%/8.7%) | Hypertension (86.3%/85%), Diabetes Mellitus (50%/40.1%), Dyslipidemia (70.4%/73.7%), History of MI (15.9%/24.1%), Previous PCI (38.6%/37.5%), Previous Stroke (9%/7.1%) | Not mentioned | Not mentioned | Not mentioned | Japan |
| Jin et al | Randomized trial | JBT-BSKT (44) & JWT (45) | MACE, cardiac death, MI, TLR, heart failure, Angina CCS>2 | 66.0 ± 8.8/65.4 ± 10.4 | STEMI (22.7%/17.8%), NSTEMI (15.9%/31.1%), UA (61.4%/51.1%), Chronic MI (9.1%/13.3%) | LAD (82.2%/75.6%), LCX (11.1%/8.9%), RCA (6.7%/15.6%) | Hypertension (56.8%/73.3%), Diabetes Mellitus (40.9%/35.6%), Dyslipidemia (25%/28.9%), History of MI (9.1%/13.3%), Previous PCI (9.1%/6.7%) | MV (mean 2.98 ± 0.36) | MV (mean 2.99 ± 0.39) | Mean 19.0 ± 6.1 month | China |
| Qu et al | Retrospective cohort | JBT-BSKT (34) & JWT (34) | SB deterioration, SB occlusion, rescue FKBI, bailout stenting, revascularization, SB loss, wire or balloon damage, MACE | 62 (IQR 13.5)/59 (IQR 10.5) | STEMI (5.9%/15.7%), NSTEMI (17.6%/23.5%), UA (76.5%/64.7%) | LM (17.6%/23.5%), LAD-D1 (76.5%/70.6%), LCX-OM (5.9%/2.9%) | Hypertension (64.7%/61.8%), Diabetes Mellitus (29.4%/29.4%), Dyslipidemia (38.2%/29.4%), History of MI (5.9%/2.9%), Previous PCI (14.7%/8.8%) | Not mentioned | Not mentioned | 12 months | China |
| Dou et al | Randomized trial | JBT (136) & JWT (151) | SB occlusion, absence of blood flow, TIMI flow grade decrease, MI | 60.8 ± 9.3/60.7 ± 10.1 | UA (63.1%/54.5%) | LAD-D1 (84.5%/81.4%), LCX-OM (13.1%/15%), RCA-PDA (2.4%/3.6%) | Hypertension (57.1%/65.3%), Diabetes Mellitus (26.2%/31.7%), Dyslipidemia (39.9%/42.5%), History of MI (21.4%/21.6%), Previous PCI (14.3%/17.4%) | MV (mean 3.0 ± 0.37) | MV (mean 2.96 ± 0.34) | 1 month | China |
| Wang et al | Randomized trial | mJBT (173) & JWT (167) | MACE, cardiac death, TLR, SB occlusion, SB rewiring | 57.01 ± 9.08/56.81 ± 8.85 | ACS (6.93%/8.38%) | LAD (78%/73%), LCX (14.45%/18.56%), RCA (7.51%/8.38%) | Hypertension (54.3%/60.4%), Diabetes Mellitus (30.6%/36.5%), Dyslipidemia (95.3%/95.8%), History of MI (6.9%/8.3%), Previous Stroke (7.5%/8.9%) | MV (mean 2.70 ± 0.30) | MV (mean 2.70 ± 0.31) | 12 months | China |
| Zhang et al | Randomized trial | JBT (143) & JWT (141) | SB occlusion, absence of blood flow, TIMI flow grade decrease, MI, MACE, all-cause mortality, cardiac death, TVR, TLR, stent thrombosis | 61.1 ± 9.1/60.9 ± 10.0 | UA (67.1%/53.9%) | LAD-D1 (85.3%/83%), LCX-OM (12.6%/13.5%), RCA-PDA (2.1%/3.5%) | Hypertension (56.6%/66%), Diabetes Mellitus (28%/29.8%), Dyslipidemia (39.2%/40.4%), History of MI (22.4%/23.4%), Previous PCI (14%/18.4%) | MV (mean 3.0 ± 0.4) | MV (mean 3.0 ± 0.3) | 12 months | China |
| Lin et al | Retrospective cohort | mJBT (30) & JWT (234) | SB occlusion, coronary dissection, wire entrapment, balloon rupture or entrapment | 61 (Q1-Q3 58-77)/63 (Q1-Q3 52-72) | STEMI (100%/99%), NSTEMI (0%/1%) | LM (0%/1.4%), LAD (93.3%/75.6%), LCX (6.7%/14.1%), RCA (0%/8.5%) | Hypertension (73.3%/61.1%), Diabetes Mellitus (53.3%/30.3%), Dyslipidemia (73.3%/68.4%), History of MI (0%/6.4%), Previous PCI (0%/8.5%), Previous Stroke (20%/9.8%) | MV (median 3.00, Q1-Q3 2.50-3.50), SB (median 2.21, Q1-Q3 2.00-2.30) | MV (median 2.75, Q1-Q3 2.75-3.00), SB (median 2.17, Q1-Q3 1.94-2.40) | Not mentioned | Taiwan |
| Qiu et al | Retrospective cohort | mJBT (105) & JWT (111) | SB occlusion, coronary dissection, MI, MACE, wire entrapment, SB dropout | 64.47 ± 9.33/66.59 ± 10.3 | STEMI (11.4%/11.7%), NSTEMI (25.7%/16.2%), UA (60%/71.2%), SCAD (1%/0.9%) | LAD-D1 (78.1%/73%), LCX-OM (18.1%/22.5%), RCA-PDA (3.8%/4.5%) | Hypertension (72.4%/64%), Diabetes Mellitus (33.3%/31.5%), Dyslipidemia (16.2%/13.5%), History of MI (19%/21.6%), Previous PCI (29.5%/26.1%), Previous Stroke (20%/20.7%) | Not mentioned | China | ||
| Cui et al | Retrospective cohort | aJBT (67) & JWT (130) | SB ostium area, coronary dissection | Not mentioned | Not mentioned | LM (14.1%/16.2%), LAD-D1 (78.9%/70.6%), LCX-OM (5.6%/10.3%), RCA-PDA (1.4%/2.9%) | Not mentioned | Not mentioned | China |
ACS = acute coronary syndrome; aJBT = active JBT; BSKT = balloon-stent kissing technique; JBT = jailed balloon technique; JWT = jailed wire technique; LAD = left anterior descending; LCX = left circumflex; LM = left main; MACE = major adverse cardiovascular events; MI = myocardial infarction; mJBT = modified JBT; NSTEACS = non-ST-elevation ACS; PCI = percutaneous coronary intervention; PDA = posterior descending artery; RCA = right coronary artery; SB = side branch; STEMI = ST-elevation myocardial infarction; TIMI = thrombolysis in myocardial infarction; TLR = target-lesion revascularization; TVR = target-vessel revascularization; UAP = unstable angina pectoris.
Procedure-related adverse events
Figure 2 shows the risk of SB occlusion during procedure was lower when using JBT compared to JWT as an SB protection during bifurcation PCI (OR = 0.31, 95% CI = 0.20 to 0.48, p <0.00001, I 2 = 39%). However, there is no significant difference in coronary dissection occurrence between JBT and JWT groups (OR = 1.54, 95% CI = 0.68 to 3.48, p = 0.30, I 2 = 42%). Similar to dissection, the rate of peri-procedural MI was similar between both groups (OR = 0.68, 95% CI = 0.26 to 1.75, p = 0.42, I 2 = 51%). Among the 10 included studies, 7 reported cases of wire/balloon entrapment or fracture. Only 1 patient in the JBT group experienced balloon entrapment. Because the jailed balloon could not be retrieved smoothly, balloon dilation followed by MV postdilation was required to correct stent strut deformation. The heterogeneity test on peri-procedural MI end point shows high heterogeneity (>50%). Thus, we used the random-effect model and conducted the sensitivity analysis ( Table 2 ). The analysis using leave-one-out method revealed the study by Kuno et al was the source of heterogeneity, which reduced the I 2 from 51% to 0% by removing the study. However, the overall effect turned from nonsignificant into statistically significant results (p <0.05). Thus, this end point should be interpreted with caution.
