T-cell receptor excision circles (TREC) may be a useful surrogate marker in HIV-1 infection for evaluating the likelihood of continued clinical stability and/or the response to therapeutics, including vaccines. Analysis of TREC in SHIV and SIV models of HIV-1 infection may provide additional information concerning the utility of TREC as a marker. We measured TREC in peripheral blood mononuclear cells (PBMC) from rhesus macaques in SHIV89.6p (n = 20) and SIVmac251 (n = 11) models of HIV-1 infection. TREC were also evaluated in tissues in the SIVmac251 model at end-point. In the SHIV89.6p model, TREC in PBMC were significantly lower at 12 weeks postinfection compared to preinfection levels. The decrease in TREC correlated with the decline in CD4+ T cells (r_s = 0.496; P = 0.026), which in turn correlated inversely with serum viral loads at end-point (r_s = -0.517; P = 0.019). Macaques that controlled SHIV89.6p infection to some degree (n = 6) had higher TREC at study end-point (P = 0.017). In the SIVmac251 model, TREC in PBMC were significantly reduced after 17 months of infection (P = 0.012) despite receiving highly active antiretroviral therapy (HAART) consisting of didanosine (ddI) and (R)-9-(2-phosphonylmethoxypropyl)-adenine (PMPA) when not cycling off therapy during scheduled treatment interruptions (STI). However, macaques that received continuous hydroxyurea (HU) in addition to the HAART regimen had higher end-point TREC compared to the non-HU group (P = 0.041), and the reduction in TREC observed at end-point within the HU group was not significant. In the SIVmac251 model, TREC correlated with the percentage of CD4+ T cells (r_s = 0.426; P = 0.048) and CD4+CD28+ T cells (r_s = 0.624; P = 0.002), and inversely with CD8+ T cells (r_s = -0.622; P = 0.002), CD8+CD28- T cells (r_s = -0.516; P = 0.014), and serum viral loads (r_s = -0.627; P = 0.039). High levels of TREC were observed in the thymus, levels comparable to PBMC were seen in the lymph node, and low but detectable levels of TREC were present in bone marrow. The use of correlates of TREC as covariates in ANCOVA revealed that the decline in TREC in the SHIV 89.6p model reflected the decline in the percentage of CD4+ T-cells due to viral cytopathogenicity. In the SIVmac251 model, the decline in TREC was related to increased immune activation and proliferation due to viral replication, as reflected by decreases in percentages of CD4+CD28+ T cells and increases in CD8+ and CD8+CD28- T cells.