Cluster analysis of blood serum inflammation markers of conditionally healthy people

Abstract

Determination of inflammatory markers in blood of conventionally healthy people is of interest due to opportunity of detecting diseases at early (preclinical) stages, as well as latent forms of pathological processes. The level of inflammation may serve as an additional criterion to forming control groups in clinical and biological studies. The aim of the study is to identify some inflammatory and autoimmune markers in a group of conventionally healthy people and to conduct a cluster analysis of the data obtained. The study involved 100 apparently healthy people (without clinical signs of infections, somatic, neurological or mental diseases) aged 19 to 88 years. The levels of IL-10, TNFα, IL-6 and autoantibodies to S100b and MBP were determined in blood serum using ELISA. Enzymatic activity of leukocyte elastase (LE) and functional activity of the α1-proteinase inhibitor (α1-PI) were determined spectrophotometrically. Protease inhibitory index (PII) was calculated as the ratio of LE to α1-PI. Cluster analysis, as well as the Shapiro–Wilk, Kruskal–Wallis, and ANOVA methods were used as the main approach to statistical data processing. All the subjects were divided into three clusters, according to their immunological parameters. The selected clusters were statistically significantly different from each other, in terms of LE activity, protease-inhibitory index (PII), as well as IL-10 and TNFα levels. The indices of a distinct cluster (43% of total cohort) are most close to average indices assessed for the general sample, which gives ground to consider the values of immune indicators from this cluster as physiological norm, corresponding to the background immunity state in healthy people. Combination of immunological parameters in two other clusters (30 and 27% of the subjects, respectively) may reflect different variants of inflammatory reactions. These clusters are characterized by multidirectional changes in LE activity and protease-inhibitory index, compared to the standard values, thus suggestive for different variants of latent inflammatory reactivity, which are realized in the patients presented in these clusters. The obtained clusters did not differ by age of the subjects (p = 0.3476), which makes it possible to exclude a significant influence of age on the determined immune parameters, and by gender characteristics (p = 0.7233). The selected clusters did not differ statistically in the functional activity of α1-PI and in the level of autoantibodies to S100b and MBP. Thus, the group of conditionally healthy people is heterogeneous in terms of inflammation markers. Inflammatory reactions of varying severity were detected in about half of the cases. Probably, this may indicate the presence of a latent pathological process and requires a detailed clinical examination.