Osteoarthritis (OA) is the leading cause of musculoskeletal disability in the elderly. Insights into the biological features of OA are obtained by characterization of the molecular features by gene expression profiling using reverse transcription-quantitative PCR (RT-qPCR). However, it has recently become evident that the use of suitable reference genes is required for appropriate normalization of this technique. Here total RNA was isolated from the synovium of 18 men and 20 women who underwent total knee arthroplasty for knee OA (KOA). We validated the expression stability of 7 candidate housekeeping genes (ACTB, B2M, GAPDH, HPRT1, RPL13A, SDHA, and YWHAZ) in the synovium of KOA with 3 commonly used algorithms (geNorm, NormFinder, and BestKeeper). Additionally, we evaluated expression profiles of the steroid hormone receptor (AR, ESR1, ESR2, GR, MR, and PR) and proinflammatory cytokines (IL1B and IL6) genes in the synovium and their correlations with the risk factors of KOA, using the most and least stable housekeeping genes for comparison. Results showed that HPRT1 was the most stable gene, whereas B2M was the least stable. RT-qPCR analysis revealed sexually dimorphic expression of AR, IL1B, and IL6; intercorrelations between steroid hormone receptor expression levels and female-specific correlations of IL1B expression with ESR1 and PR expression, IL6 expression with ESR1 and GR expression, and body mass index with AR and PR expression; and the choice of the least stable reference gene altered several correlations and statistical significances. In conclusion, HPRT1 was identified as the suitable reference gene for normalization in the OA synovium.