Inflammation is a hallmark of Alzheimer's disease (AD). Whether directly involved in the pathogenesis, or a downstream consequence of neuronal death, the blood neutrophil–lymphocyte ratio (NLR) is reported to be a putative, non-invasive peripheral biomarker for AD. The aim of this study was to re-evaluate the diagnostic utility of longitudinal measures of the NLR. The NLR was stable across all time-points and weakly correlated with neocortical amyloid burden (R = 0.21 at baseline, 0.27 at 18 months, 0.20 at 36 months and 0.10 at 54 months). Cross-sectionally, the NLR was significantly elevated in AD participants as compared to HC participants at baseline (p 0.05 at all time-points except for 18 months; p 0.05) adjusted for age, sex and APOEε4 allele status. Comparing the NLR between cognitive transition groups over time (transition towards an AD type dementia), there was no significant difference in the NLR levels between those participants, who did not transition and those participants who did transition, or those in the stable AD group after adjusting for age, sex and APOEε4 allele status (p > 0.05). Despite inflammation being a hallmark in AD and previous reports showing that the NLR can discriminate HC from AD patients, our results suggest that the sensitivity of the NLR itself is not robust enough for diagnostic utility. We identified significant relationships cross sectionally (p 0.5). The NLR also had limited association with cognitive decline, although in our cohort, the number of participants transitioning was relatively small. In conclusion, the NLR may reflect AD-related inflammatory processes in the periphery, but age and sex are dominant covariates which need to be controlled for in population-based screening.