Background: Airway microbiota dynamics during lower respiratory infection (LRI) are still poorly understood due, in part, to insufficient longitudinal studies and lack of uncontaminated lower airways samples. Furthermore, the similarity between upper and lower airway microbiomes is still under debate. Here we compare the diversity and temporal dynamics of microbiotas directly sampled from the trachea via tracheostomy in patients with (YLRI) and without (NLRI) lower respiratory infections. Methods: We prospectively collected 127 tracheal aspirates across four consecutive meteorological seasons (quarters) from 40 patients, of whom 20 developed LRIs and 20 remained healthy. All aspirates were collected when patients had no LRI. We generated 16S rRNA-based microbial profiles (~250 bp) in a MiSeq platform and analyzed them using Mothur and the SILVAv123 database. Differences in microbial diversity and taxon normalized (via negative binomial distribution) abundances were assessed using linear mixed effects models and multivariate analysis of variance. Results and discussion: Alpha-diversity (ACE, Fisher and phylogenetic diversity) and beta-diversity (Bray-Curtis, Jaccard and Unifrac distances) indices varied significantly (P<0.05) between NLRI and YLRI microbiotas from tracheostomised patients. Additionally, Haemophilus was significantly (P = 0.009) more abundant in YLRI patients than in NLRI patients, while Acinetobacter, Corynebacterium and Pseudomonas (P<0.05) showed the inverse relationship. We did not detect significant differences in diversity and bacterial abundance among seasons. This result disagrees with previous evidence suggesting seasonal variation in airway microbiotas. Further study is needed to address the interaction between microbes and LRI during times of health and disease. © 2017 Pérez-Losada et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.