BACKGROUND: The aim of this study was to investigate alterations in homocysteine levels with delays in the initiation time of centrifugation. Attempts were made to improve the processing of samples by analyzing the homocysteine levels of patients from a health promotion center, where delays in the initiation of sample centrifugation were expected because of the nature of the workload and operational procedures. METHODS: Forty healthy adults were selected for the measurement of homocysteine levels. The two samples those were obtained from each individual simultaneously and separately were designated as group A (centrifugation initiation time < or =20 minutes after blood sampling) and group B (> or =2 hours). The degree of deviation from the homocysteine reference interval was also evaluated with samples from 1,134 adults who had medical check-up at a health promotion center from August 1, 2009 to July 31, 2010. RESULTS: The mean serum homocysteine level in group B was significantly higher than that of group A (12.4 vs. 10.6 micromol/L, P<0.0001). Homocysteine level increased with a mean of 19.4% when the initiation of centrifugation was delayed over 2 hours. In the 1,134 adults who had medical check-up, the numbers outside the reference interval were 334 (29.5%). Abnormal samples with homocysteine levels out of the reference interval were reduced to about 7% in July 2011 through improving sample processing to minimize delays in centrifugation initiation time. CONCLUSIONS: Homocysteine levels rapidly increase as blood sample centrifugation is delayed. Therefore, in order to provide accurate test results immediate centrifugation of blood samples is critical.