Background Rapid diagnostic tests (RDTs) detecting the histidine-rich protein 2 (PfHRP2) have a central position for the management of Plasmodium falciparum infections. Yet, variable detection of certain targeted motifs, low parasitaemia, but also deletion of pfhrp2 gene or its homologue pfhrp3, may result in false-negative RDT leading to misdiagnosis and delayed treatment. This study aimed at investigating the prevalence, and understanding the possible causes, of P. falciparum RDT-negative infections at Montpellier Academic Hospital, France. Methods The prevalence of falsely-negative RDT results reported before and after the introduction of a loop-mediated isothermal amplification (LAMP) assay, as part as the malaria screening strategy in January 2017, was analysed. Negative P. falciparum RDT infections were screened for pfhrp2 or pfhrp3 deletion; and exons 2 were sequenced to show a putative genetic diversity impairing PfHRP2 detection. Results The overall prevalence of P. falciparum negative RDTs from January 2006 to December 2018 was low (3/446). Whereas no cases were reported from 2006 to 2016 (0/373), period during which the malaria diagnostic screen was based on microscopy and RDT, prevalence increased up to 4.1% (3/73) between 2017 and 2018, when molecular detection was implemented for primary screening. Neither pfhrp2/3 deletion nor major variation in the frequency of repetitive epitopes could explain these false-negative RDT results. Conclusion This paper demonstrates the presence of pfhrp2 and pfhrp3 genes in three P. falciparum RDT-negative infections and reviews the possible reasons for non-detection of HRP2/3 antigens in a non-endemic setting. It highlights the emergence of falsely negative rapid diagnostic tests in a non-endemic setting and draws attention on the risk of missing malaria cases with low parasitaemia infections using the RDT plus microscopy-based strategy currently recommended by French authorities. The relevance of a novel diagnostic scheme based upon a LAMP assay is discussed.