From June 1, 2009 to June 30, 2011, patients suffering from rheumatic heart disease and scheduled for valve replacement under cardiopulmonary bypass were eligible for this study. Only the cases of the first day were included, because EPCs must be processed within 2 hours after harvest. Three hundred and twenty-one patients were screened for the Endothelial Progenitor Cell and Prognosis after Cardiac Surgery study (EPCPCS). One hundred and eighty-three patients were excluded because of pulmonary insufficiency (N=47), severe pulmonary artery hypertension (N=21), chronic obstructive pulmonary disease (N=14), asthma (N=4), New York Heart Association (NYHA) classification of IV (N=38), renal insufficiency (N=12), hepatic insufficiency (N=6), repeat surgery for valve replacement (N=5), or refusal to sign the consent (N=36). Therefore, 138 patients were included in the study. All the patients signed informed consents before their participation. This Cohort study was approved by the Ethics Committee of Sichuan University, and registered with the Chinese Clinical Trial Registry (No. ChiCTR-OCS-09000398; www.chictr.org/).
Anesthesia, Cardiopulmonary Bypass and Surgery
All patients received anesthesia and cardiopulmonary bypass according to the clinical practice in West China Hospital (14). Simply, anesthesia was achieved with fentanyl, midazolam and nondepolarizing muscle relaxants, and maintained with propofol and/or inhalation agents. CPB was conducted with a roller pump (Sarns 8000, 3M) and membrane oxygenator (Medtronic Inc, Minneapolis, MN) primed with 500 ml multiple electrolytes injection (Baxter), 1000 ml hydroxyethyl starch (Fresenius Kabi) and 3750 U heparin (porcine; Tianjin Biochem Pharmaceutical Co, Ltd). During surgery, hemodynamic and fluid management were performed according to routine clinical practice in our hospital, based on central venous pressure (CVP), blood pressure and clinical evaluation of the attending anesthesiologist. After termination of CPB, residual pump blood was collected in a bag containing sodium citrate, neutralized by protamine, and returned to the patient. In Cardiac Intensive Care Unit (ICU), patients received analgesia based on the clinical evaluation of the attending doctor in ICU.
Preparation of Blood Samples
For flow cytometry and plasma factors determination, blood samples were drawn before 4 and 20 hours after surgery from a central venous catheter, anti-coagulated with heparin, and centrifuged for 15 min at 1000 g and 4 ℃. The plasma was stored at -80 ℃ until analysis. Both G-CSF and Vascular Endothelial Growth Factor (VEGF) levels were measured by Enzyme Linked Immunoadsorbent Assay kits (R&D Systemics, Inc, Minneapolis, MN and Bender MedSystems GmbH, Vienna Austria), according to the manufacturers' instruction. For the EPC counts, central venous blood was collected into a Vacutainer tube (Becton Dickinson, Basel, Switzerland) with sodium citrate anticoagulant before and after surgery.
After the plasma was removed, red blood cells were lysed with ammonium chloride solution. For fluorescence-activated cell-sorting analysis, cells were re-suspended in 100.0 μL of phosphate-buffered saline (PBS). Nonspecific antibody binding was blocked using 20.0 μL FcR-blocking reagent (130-059-901, Miltenyi Biotec) for 20 minutes at room temperature before staining with conjugated antibodies. Immunofluorescent cell staining was performed with 6.0 μL of CD34-FITC (Becton Dickinson), 8.0 μL of CD309-PerCP/Cy5.5 (KDR, BioLegend) and 6.0 μL of CD133/1(AC133)-APC (Miltenyi Biotec). Each antibody was titrated by serial dilutions. IgG1-FITC/PE/APC PerCP/Cy5.5 antibody (Becton Dickinson) served as a negative control. These surface antibodies were incubated for 30 min at room temperature in the dark, followed by staining by DAPI (Sigma Aldrich) for dead cells (15-17). Data acquisition was performed on a FACS Aria cytometer equipped with FACS Diva 5.0 software (BD) and analyzed by Flowjo software (Tree Star). The instrument setup was standardized to reduce batch-to-batch differences by daily monitoring with Rainbow Beads (Becton Dickinson). A minimum of 5,000,000 events was acquired. The boundary between positive and negative cells was placed using fluorescence-minus-one controls and an internal control (16, 17). Figure 1 illustrates the sequential gating strategy to mark EPCs for intensive analysis (17). For each patient, a corresponding negative control with IgG1-FITC/PerCP/Cy5.5/APC was obtained. The numbers of CD34+KDR+ and CD34+KDR+ CD133+ EPCs were derived from the absolute number of white blood cells (WBCs) provided by a hematological analyzer (XE-5000, Sysmex, Kobe, Japan) and the percentages of CD34+KDR+ and CD133+CD34+KDR+EPCs were determined by flow cytometry, using the following formula: percentage of EPCs×WBC count/100 (18).
The primary outcomes were ARDS, renal dysfunction and heart failure. The arterial partial pressure of oxygen (PaO2) and fraction of inspired oxygen (FiO2) were measured before 4 and 20 hours after surgery. ARDS was defined as respiratory failure without cardiac dysfunction or fluid overload at 20 hours after surgery, which was objectively assessed by echocardiography, according to the Berlin definition of ARDS (19). Mild ARDS was defined as PaO2/FiO2 between 200-300 mm Hg with PEEP or CPAP≥5 cm H2O; moderate as 100 mm Hg ≤ PaO2/FiO2 ≤ 200 mm Hg with PEEP ≥5 cm H2O, and severe as PaO2/FiO2≤ 100 mm Hg with PEEP ≥5 cm H2O. Renal dysfunction was defined as a postoperative serum creatinine level over 177 μM with an increase of more than 62 μM; renal failure was defined as requiring dialysis or in-hospital death associated with acute renal dysfunction (20). Heart failure was defined as sudden deaths, congestive heart failure, and acute myocardial infarction.
Data were analyzed by SAS 9.1 statistical software (SAS Institute Inc., Cary, NC). All quantitative data were examined for their distribution. Correlation analysis was used to examine the relationship between variables. Pearson and Spearman correlation analyses were used for normal and non-normal distribution data respectively.
Normally distributed variables including age, body mass index, and blood cells were expressed as mean ± standard deviation, and one-way ANOVA was used to compare the difference among groups. Otherwise, non-normal distribution data including CPB, cross-clamp, and surgery times were expressed as the median and 95% confidence interval, and differences between groups were assessed with the Kruskal-Wallis. Segmental data including smoking, hypertension, diabetes, NYHA, type of surgery, death, ARDS, acute renal dysfunction, heart failure and composite outcome event were expressed as a percentage, and differences between groups were compared using chi-square or Fisher exact tests. To analyze the impact of EPC homing on outcomes, we used pre-specified thresholds corresponding to patients' decreased EPC (low, medium, and high) at 4 hours after surgery. Multivariable Logistic Regression was used for correction of the outcome between groups by age, gender, body-mass index, smoking, hypertension, diabetes, and the NYHA Functional Classification. A nonparametric signed-rank test was used to compare differences of EPC counts and G-CSF, VEGF levels at different time points within the same group. P<0.05 was considered statistically significant.