Serum levels of tumor necrosis factor-alpha (TNF-alpha), interleukin 6 (IL-6), and soluble interleukin 6 receptor (sIL-6R) have been studied as risk factors of cardiovascular disease in longitudinal studies. However, it is unknown about their long-term intra-individual variations and whether single measurements of these cytokines and receptor are reliable biomarkers in epidemiological studies. In this study, serum levels of TNF-alpha, IL-6, and sIL-6R were assayed by ELISAs in 36 young, healthy women from whom three blood samples were collected at 12-month intervals over 2 years, and the intraclass correlation coefficients (ICC) were estimated. The ICC of 0.73 (95% CI=0.49-0.79) for TNF-alpha was comparable to those of other commonly used biomarkers, justifying its use in epidemiological studies. The ICC of 0.48 (95% CI=0.25-0.58) for IL-6 was not optimal. However, IL-6 has been demonstrated as a consistent risk factor for cardiovascular disease, suggesting it could still be a useful biomarker if its disease association is substantial. The ICC of 0.36 (95% CI=0.10-0.47) for sIL-6R was relatively low, and multiple samples would need to be collected in prospective studies for this receptor.