Abstract
Background: With advancing age, the composition of leukocyte subpopulations in peripheral blood is known to change, but how this change differs between men and women and how it relates to frailty is poorly understood. Thus, our aim in this exploratory study was to investigate whether frailty is associated with changes in immune cell subpopulations and whether associations differ between men and women. Therefore, we performed in-depth immune cell phenotyping by enumerating a total of 37 subsets of T cells, B cells, NK cells, monocytes, and neutrophils in peripheral blood of 289 elderly people between 60-87 years of age. Associations between frailty and each immune cell subpopulation were tested separately in men and women and were adjusted for age and CMV serostatus. In addition, a random forest algorithm was used to predict a participant’s frailty score based on enumeration of immune cell subpopulations. Results: In an association study, frailty was observed to be associated with increases in numbers of neutrophils in both men and in women. Furthermore, sex-specific associations were found. Frailer women, but not men, showed higher numbers of total and CD16^-^ monocytes and lower numbers of CD56^+^ T cells. Overall, the accuracy of the predictions in the random forest analysis was low (9.2% explained variance in men and 12.2% in women). Yet, the random forest analysis confirmed all associations mentioned above, but did not confirm a possible negative association in women between frailty and late differentiated CD4^+^ TemRA cells. Moreover, the random forest analysis revealed additional relationships with frailty, with frailer men showing higher CD16^+^ monocyte and lower naïve B cell numbers. Other important variables for predicting frailty were plasmablast numbers in men and total T cell numbers in women. Conclusions: We report on observed associations of frailty with elevated myeloid cell numbers in men and women. In-depth immune cellular profiling also revealed sex-specific associations of frailty with several immune subpopulations. However, an expected positive association between frailty and memory T cells was not observed. We hope that our study will prompt further investigation into the immune mechanisms associated with the development of frailty in men and women.
Random forest analysis in identifying the importance of obesity risk factors
