Background: Estimation of glucose is very important in clinical diagnosis of diabetes mellitus, hyperglycemia and hypoglycemia. The core aim of clinical laboratory service is to produce accurate, precise and reliable test results. The ability of laboratories to do this lies in the availability of well trained personnel, equipment, electricity etc. There has been a progressive increase in the prevalence of diabetes mellitus in Nigeria and the burden is expected to increase even further and the need for proper diagnosis cannot be overemphasized.
 Aim: In this study we have tried to evaluate the impact of delayed separation of plasma and serum from cells and clot on glucose level.
 Method: A total of fifty (50) randomly selected apparently healthy students of Federal School of Medical Laboratory Technology (Science), Jos were recruited as subject into this study. Blood samples were collected from each of the subjects into plain and anticoagulant bottles for serum and plasma respectively. An aliquot was analyzed within 10 minutes to obtain a baseline value where other values after timely delays were compared. The samples were spun but not separated from the cells/clot all through the period of the analysis. Blood glucose was determined by glucose Oxidase Colourimetric assay kit obtained from Randox Laboratories Limited United Kingdom. Data were analyzed using student’s t-test and performed using the Statistical Package for Social Sciences (SPSS) version 20.0.
 Results: Result obtained shows the mean and standard deviation of all timing for plasma and serum. In plasma it shows that the means±SD of 2, 4 and 24hours (5.2±0.7, 5.1±0.4 and 5.3±0.5 respectively) were not significantly varied when compared to the baseline (0 hour) value (5.3±0.5). That of baseline and 24 hours (5.3±0.5 and5.3±0.5 respectively) were significantly higher (P<0.001) than that of 48 and 74 hours (1.0±0.20 and 7±0.2 respectively). In serum, similar results were obtained.
 Conclusion: In conclusion, where possible, it is advised that the established best practices in processing and analysis of samples be adhered to. However, the results of this work shows reasonably stable results for blood glucose determination could be obtained within 24 hours. Therefore, instead of discarding the sample, it might seem more appropriate that the sample be analyzed and the result reported indicating the number of hours delayed within 24hours.