BACKGROUND The heat-labile nature of Dengue virus (DENV) in serum samples must be considered when applying routine diagnostic tests to avoid issues that could impact the accuracy of test results with direct implications for case management and disease reporting. OBJECTIVES To check if pre-analytical variables, such as storage time and temperature, have an impact on the accuracy of the main routine diagnostic tests for dengue. METHODS Virus isolation, reverse transcription real-time polymerase chain reaction (RT-PCR) and NS1 enzyme-linked immunosorbent assay (ELISA) were evaluated using 84 samples submitted to different pre-analytical conditions. FINDINGS Sensitivity and negative predictive value were directly affected by sample storage conditions. RT-PCR and virus isolation showed greater dependence on well-conserved samples for an accurate diagnosis. Interestingly, even storage at -30ºC for a relatively short time (15 days) was not adequate for accurate results using virus isolation and RT-PCR tests. On the other hand, NS1 ELISA showed no significant reduction in positivity for aliquots tested under the same conditions as in the previous tests. MAIN CONCLUSIONS Our results support the stability of the NS1 marker in ELISA diagnosis and indicate that the accuracy of routine tests such as virus isolation and RT-PCR is significantly affected by inadequate transport and storage conditions of serum samples.