Ischemic Preconditioning Effect on Myocardial and Renal Injury
Ischemic Preconditioning Effect on Myocardial and Renal Injury
The meta-analysis was conducted in accordance with the QUOROM guidelines. A comprehensive and systematic search of PUBMED (1950–2010), Cochrane Library, Scielo, EMBASE, SCOPUS, and ISI Web of Science databases was performed using the following search terms: "remote ischemic preconditioning" OR "remote ischaemic preconditioning" OR "RIPC" OR "ischemic preconditioning" OR "ischaemic preconditioning" AND "heart surgery" OR "cardio protective surgery" OR "myocardial surgery" OR "heart infarction protection surgery" AND "human" OR "patient" OR "individual" OR "subject" OR "men" OR "man" OR "women" OR "woman." Two authors independently selected studies based on their titles and abstracts. In the second evaluation, both authors reviewed the complete versions of all articles and included or excluded them after a common consensus. Finally, we sought cross references from relevant articles and tried to contact corresponding authors in order to obtain unpublished data.
Studies that were included met the following criteria: prospective randomized controlled trials of patients undergoing cardiac surgery for coronary artery disease or valve replacement surgery, PCI (elective or primary), or AAA repair that had at least one arm randomized to remote ischemic preconditioning and another to placebo. Other interventions were only accepted if they were equally distributed to all arms of the study. Studies were excluded if they were duplicate publications or reported no data on rate of myocardial infarction, levels of CK-MB, troponin T, troponin I, or creatinine. There was no restriction on language. Data were abstracted from the studies not only for the cardiac and renal biomarkers, but also for clinical outcomes, including the incidence of myocardial infarction, stroke/cerebrovascular accident, atrial fibrillation, ventricular arrhythmias, congestive heart failure, inotropic use, need for urgent hemodialysis, and mortality. The clinical outcomes were determined by preset criteria for original studies.
Since 24-hour postoperative troponin levels are strongly related to short- and long-term outcomes in patients undergoing cardiovascular interventions, we used this time interval for our analysis. When this information was not readily available, we contacted the authors; if we did not receive a response, CK-MB/troponin levels at 24 hours, peak levels, or area under the curve over 72 hours were used for the analysis.
Different markers of myocardial necrosis and creatinine levels had to be corrected to standard mean differences (SMD) to allow for comparison between different studies. As a general rule, a pooled SMD lower than 0.4 is considered small and greater than 0.7 is considered large. For the articles that reported medians and interquartile ranges, means and SMDs were estimated using the algorithm proposed by Hozo et al.
The definition of clinical endpoints, such as "myocardial infarction," was predetermined by the authors of original studies. For creatinine levels, the highest value in the first 72 hours after surgery was used. Myocardial infarction was reported as rate of events and summarized as a risk ratio. Heterogeneity was evaluated using the I2 test. If results were considered homogenous, a fixed effects model was used for data synthesis; if not, a random effects model was used. Publication bias was evaluated by visual analysis of the funnel plot. Alpha was set at P<.05. Data analysis was done using Review Manager 5 Software (2010, The Cochrane Collaboration).
The following subgroup analyses were planned: type of surgery, type of ischemic stimulus, use of beta-blockers, diabetes, and use of sevoflurane or desflurane. Sensitivity analysis was conducted in all comparisons with significant heterogeneity, by systematic elimination of each study from the analysis, until results were as homogenous as possible.
Methods
Search Strategy and Data Extraction
The meta-analysis was conducted in accordance with the QUOROM guidelines. A comprehensive and systematic search of PUBMED (1950–2010), Cochrane Library, Scielo, EMBASE, SCOPUS, and ISI Web of Science databases was performed using the following search terms: "remote ischemic preconditioning" OR "remote ischaemic preconditioning" OR "RIPC" OR "ischemic preconditioning" OR "ischaemic preconditioning" AND "heart surgery" OR "cardio protective surgery" OR "myocardial surgery" OR "heart infarction protection surgery" AND "human" OR "patient" OR "individual" OR "subject" OR "men" OR "man" OR "women" OR "woman." Two authors independently selected studies based on their titles and abstracts. In the second evaluation, both authors reviewed the complete versions of all articles and included or excluded them after a common consensus. Finally, we sought cross references from relevant articles and tried to contact corresponding authors in order to obtain unpublished data.
Studies that were included met the following criteria: prospective randomized controlled trials of patients undergoing cardiac surgery for coronary artery disease or valve replacement surgery, PCI (elective or primary), or AAA repair that had at least one arm randomized to remote ischemic preconditioning and another to placebo. Other interventions were only accepted if they were equally distributed to all arms of the study. Studies were excluded if they were duplicate publications or reported no data on rate of myocardial infarction, levels of CK-MB, troponin T, troponin I, or creatinine. There was no restriction on language. Data were abstracted from the studies not only for the cardiac and renal biomarkers, but also for clinical outcomes, including the incidence of myocardial infarction, stroke/cerebrovascular accident, atrial fibrillation, ventricular arrhythmias, congestive heart failure, inotropic use, need for urgent hemodialysis, and mortality. The clinical outcomes were determined by preset criteria for original studies.
Statistical Analysis
Since 24-hour postoperative troponin levels are strongly related to short- and long-term outcomes in patients undergoing cardiovascular interventions, we used this time interval for our analysis. When this information was not readily available, we contacted the authors; if we did not receive a response, CK-MB/troponin levels at 24 hours, peak levels, or area under the curve over 72 hours were used for the analysis.
Different markers of myocardial necrosis and creatinine levels had to be corrected to standard mean differences (SMD) to allow for comparison between different studies. As a general rule, a pooled SMD lower than 0.4 is considered small and greater than 0.7 is considered large. For the articles that reported medians and interquartile ranges, means and SMDs were estimated using the algorithm proposed by Hozo et al.
The definition of clinical endpoints, such as "myocardial infarction," was predetermined by the authors of original studies. For creatinine levels, the highest value in the first 72 hours after surgery was used. Myocardial infarction was reported as rate of events and summarized as a risk ratio. Heterogeneity was evaluated using the I2 test. If results were considered homogenous, a fixed effects model was used for data synthesis; if not, a random effects model was used. Publication bias was evaluated by visual analysis of the funnel plot. Alpha was set at P<.05. Data analysis was done using Review Manager 5 Software (2010, The Cochrane Collaboration).
The following subgroup analyses were planned: type of surgery, type of ischemic stimulus, use of beta-blockers, diabetes, and use of sevoflurane or desflurane. Sensitivity analysis was conducted in all comparisons with significant heterogeneity, by systematic elimination of each study from the analysis, until results were as homogenous as possible.
