Centrifugal pump and roller pump in adult cardiac surgery: a meta-analysis of randomized controlled trials.

Centrifugal pump (CP) and roller pump (RP) designs are the dominant main arterial pumps used in cardiopulmonary bypass (CPB). Trials reporting clinical outcome measures comparing CP and RP are controversial. Therefore, a meta-analysis was undertaken to evaluate clinical variables from randomized controlled trials (RCTs). Keyword searches were performed on Medline (1966-2011), EmBase (1980-2011), and CINAHL (1981-2011) for studies comparing RP and CP as the main arterial pump in adult CPB. Pooled fixed-effects estimates for dichotomous and continuous data were calculated as an odds ratio and weighted-mean difference, respectively. The P value was utilized to assess statistical significance (P < 0.05) between CP and RP groups. Eighteen RCTs met inclusion criteria, which represented 1868 patients (CP = 961, RP = 907). The prevailing operation was isolated coronary artery bypass graft surgery (CP = 88%, RP = 87%). Fixed-effects pooled estimates were performed for end-of-CPB (ECP) and postoperative day one (PDO) for platelet count (ECP: P = 0.51, PDO: P = 0.16), plasma free hemoglobin (ECP: P = 0.36, PDO: P = 0.24), white blood cell count (ECP: P = 0.21, PDO: P = 0.66), and hematocrit (ECP: P = 0.06, PDO: P = 0.51). No difference was demonstrated for postoperative blood loss (P = 0.65) or red blood cell transfusion (P = 0.71). Intensive care unit length of stay (P = 0.30), hospital length of stay (P = 0.33), and mortality (P = 0.91) were similar between the CP and RP groups. Neurologic outcomes were not amenable to pooled analysis; nevertheless, the results were inconclusive. There was no reported pump-related malfunction or mishap. The meta-analysis of RCTs comparing CP and RP in adult cardiac surgery suggests no significant difference for hematological variables, postoperative blood loss, transfusions, neurological outcomes, or mortality.