Stent Loss and Retrieval During PCI
Stent Loss and Retrieval During PCI
We performed a search of the Pubmed and Cochrane Library databases for manuscripts on coronary stent loss and stent retrieval. Bibliographies of the retrieved studies were searched by hand for other relevant studies. Human studies in English published between the years 1991 to 2012 were included if they reported the frequency of stent loss and associated treatment (including retrieval or deployment attempts) and clinical outcomes. Stent loss was defined as dislodgment of the stent from the stent delivery balloon in a location different that the intended area for deployment. Review articles were excluded, but case reports were included. A list of the included studies is shown in Supplemental Table 1 (available at www.invasivecardiology.com).
Data from included studies were extracted by two authors (MA and TM) and controversies were reviewed by both. In case of disagreement, a third reviewer (EB) was used to reach consensus. Extracted data included the total number of patients, the frequency of stent loss, the various treatment strategies (retrieval vs deployment vs crushing), age, sex, history of prior coronary artery bypass graft (CABG), target vessel, mechanisms of loss, clinical presentation, treatment including the use of retrieval devices/techniques, and occurrence of complications. The country and institute of origin, author, and enrollment period were reviewed to identify and exclude duplicate publications from the same cohort. The authors of 8 studies that reported incomplete information were contacted and two of them provided additional data for the present meta-analysis. To minimize the impact of publication bias inherent in case reports, only data reported in case series were used to estimate the frequency of stent loss and the resultant complications (ie, case reports were excluded from this part of the analysis).
The frequency of each evaluated outcome was abstracted from each study and presented as minimum, maximum, and cumulative rates. To assess heterogeneity across trials, we used the Cochrane Q statistic (heterogeneity P-value ≤.10 was considered significant) and I statistic (25%, 50%, and 75% correlate with low, moderate, and high heterogeneity, respectively) for each outcome. Due to its conservative or "worst-case scenario" estimates, a random-effects model as described by DerSimonian and Laird was used to obtain a summary estimate and 95% confidence intervals (CIs). The cumulative proportion and corresponding CIs are presented in Supplemental Table 2 (www.invasivecardiology.com). Data collection, study selection, processing of the data, and reporting of the results were performed according to accepted principles related to systematic review and meta-analysis. Publication bias for the frequency of stent loss and the occurrence of complications (death, emergency CABG, and composite major adverse cardiac events [MACE] endpoint) were estimated visually by funnel plots and are presented in Supplemental Figures 1, 2, 3, and 4 (www.invasivecardiology.com). Proportions were compared using the chi-square test and P<.05 was considered statistically significant.
(Enlarge Image)
Supplemental Figure 1.
Funnel plot for stent loss rate.
(Enlarge Image)
Supplemental Figure 2.
Funnel plot for the incidence of major adverse cardiac events.
(Enlarge Image)
Supplemental Figure 3.
Funnel plot for the frequency of coronary artery bypass graft surgery.
(Enlarge Image)
Supplemental Figure 4.
Funnel plot for the incidence of death.
Methods
Search Strategy and Eligibility Criteria
We performed a search of the Pubmed and Cochrane Library databases for manuscripts on coronary stent loss and stent retrieval. Bibliographies of the retrieved studies were searched by hand for other relevant studies. Human studies in English published between the years 1991 to 2012 were included if they reported the frequency of stent loss and associated treatment (including retrieval or deployment attempts) and clinical outcomes. Stent loss was defined as dislodgment of the stent from the stent delivery balloon in a location different that the intended area for deployment. Review articles were excluded, but case reports were included. A list of the included studies is shown in Supplemental Table 1 (available at www.invasivecardiology.com).
Data Extraction
Data from included studies were extracted by two authors (MA and TM) and controversies were reviewed by both. In case of disagreement, a third reviewer (EB) was used to reach consensus. Extracted data included the total number of patients, the frequency of stent loss, the various treatment strategies (retrieval vs deployment vs crushing), age, sex, history of prior coronary artery bypass graft (CABG), target vessel, mechanisms of loss, clinical presentation, treatment including the use of retrieval devices/techniques, and occurrence of complications. The country and institute of origin, author, and enrollment period were reviewed to identify and exclude duplicate publications from the same cohort. The authors of 8 studies that reported incomplete information were contacted and two of them provided additional data for the present meta-analysis. To minimize the impact of publication bias inherent in case reports, only data reported in case series were used to estimate the frequency of stent loss and the resultant complications (ie, case reports were excluded from this part of the analysis).
Statistical Analysis
The frequency of each evaluated outcome was abstracted from each study and presented as minimum, maximum, and cumulative rates. To assess heterogeneity across trials, we used the Cochrane Q statistic (heterogeneity P-value ≤.10 was considered significant) and I statistic (25%, 50%, and 75% correlate with low, moderate, and high heterogeneity, respectively) for each outcome. Due to its conservative or "worst-case scenario" estimates, a random-effects model as described by DerSimonian and Laird was used to obtain a summary estimate and 95% confidence intervals (CIs). The cumulative proportion and corresponding CIs are presented in Supplemental Table 2 (www.invasivecardiology.com). Data collection, study selection, processing of the data, and reporting of the results were performed according to accepted principles related to systematic review and meta-analysis. Publication bias for the frequency of stent loss and the occurrence of complications (death, emergency CABG, and composite major adverse cardiac events [MACE] endpoint) were estimated visually by funnel plots and are presented in Supplemental Figures 1, 2, 3, and 4 (www.invasivecardiology.com). Proportions were compared using the chi-square test and P<.05 was considered statistically significant.
(Enlarge Image)
Supplemental Figure 1.
Funnel plot for stent loss rate.
(Enlarge Image)
Supplemental Figure 2.
Funnel plot for the incidence of major adverse cardiac events.
(Enlarge Image)
Supplemental Figure 3.
Funnel plot for the frequency of coronary artery bypass graft surgery.
(Enlarge Image)
Supplemental Figure 4.
Funnel plot for the incidence of death.
Source...