OBJECTIVES:  In laboratory diagnostics, 70% of errors occur in the preanalytical phase. Our purpose was to identify factors that increase hemolysis in blood samples collected from intravenous catheters and to examine whether incorrect order of draw results in biased test results. METHODS:  This quantitative, cross-sectional study was carried out in Karolina Hospital, Mosonmagyaróvár, Hungary between September 1. and November 30, 2016. Volunteers were selected by non-randomized purposive sampling. Blood was collected from 120 volunteers using intravenous catheters of various sizes (18, 20, 22 and 24G). During phlebotomy the following parameters were recorded: catheter size, tourniquet time, difficulty of phlebotomy, blood flow through the cannula, site of cannula insertion. Data were analyzed using descriptive (absolute and relative frequency, median, interquartile range – IQR) and mathematical (χ2-test, Mann- Whitney U tests, linear and logistic regressions) statistical analyses (p <0.05). RESULTS:  The rate of hemolyzed samples was 13.3% (16/120). Hemolysis was significantly increased by the use of intravenous cannula 22-24G (OR: 3.638, 95% CI: 1.101-12.027, p <0.05), tourniquet time over one minute (OR: 4.256, 95% CI: 1.286-14.087, p <0.05) and site of blood draw distal to the antecubital fossa (OR: 6.143, 95% CI: 2.023-18.653, p<0,05). Increased tourniquet time resulted in increased LDH (r=0.226, p=0.013) and CK (r=0.248, p=0.006) levels. CONCLUSIONS:  Using cannula >20G, tourniquet time longer than 1 minute and drawing blood from sites other than the antecubital fossa result in increased risk of hemolysis. Increased toniquet time also resulted in the increase of certain analytes.