Septic shock is the lethal outcome of an infectious process. In developing countries, such as Colombia, rates of septic shock are very high (F. Rodríguez et al 2011, Angus DC et al 2001), leading to human lost and economic costs. Unfortunately, to address this issue, we lack a reliable, sensitive and specific predictor of septic shock. Hence, identifying such a predictor is a cornerstone to deliver early therapy, therefore decreasing mortality (Rivers E, et al. 2001). Here, we aimed to evaluate the reliability of the so-called inflammatory reflex (focusing on inflammatory response and autonomic dysfunction) as predictor of septic shock. Over a period of 6 months, we conducted a prospective study in adults admitted to the ICU of Hospital San Ignacio, Bogota, Colombia, who met criteria for sepsis (Levy MM et al 2003). Appropriate written consents were obtained. Besides standard analyses, a single continuous 15 minute ECG was recorded using a digital polygraph (LabChart® v7.1, HRV Module®, ADInstruments®) and measures of HRV were calculated. Levels of Epinephrine, Norepinephrine, TNF and Interleukins were obtained from blood samples by chromatography or fluorometry. Blood chemistry and HRV were correlated with outcome (shock vs. no shock) after 24 hours. Statistical analyses were performed using PAST (v. 3.01). Regarding HRV, we observed an increase in RMSSD (Student-test, t11,22= 2.59, p<0.01; mean±SD: shock= 165.96 ms ± 182.4; no-shock= 50.10 ms ± 70.02) and the HF component (Student-test, t11,22= 2.72, p<0.01; mean±SD: shock= 11742.1ms2 ± 1963.87 ms2; no-shock= 777.38 ± 1513.2 ms2) in the baseline measurements of patients who developed shock. In contrast, we observed significantly lower levels of LF (Student-test, t11,22= 2.12, p<0.05; mean±SD: shock= 4086.62 ± 556.62 ms2; no-shock= 1245.93 ± 215.73 ms2) and LF/HF ratio (Student-test, t11,22= 2.69, p<0.01, mean±SD: shock= 0.46 ms2 ± 0.28; no-shock= 4.68 ms2 ± 5.15) in the baseline measurement of patients who developed shock. Moreover, patients exhibited lower levels of IL-6 (Student-test, t9,20= 2.11, p<0.05, mean±SD: shock= 133.35 pg/ml ± 140.7; no-shock= 876.17 pg/ml ± 1038.45) in the baseline measurements of patients who developed shock. We did not observed significant differences across other variables. Our results indicate that patients who progress to shock have an increased inflammatory reflex (altered autonomic response with a predominantly vagal activity and a decreased sympathetic-vagal ratio with a decreased inflammatory response). We conclude that specific components of the inflammatory reflex, namely HRV and IL-6, are reliable predictors of septic shock in critical patients. In the mid term we will aim to try using these predictors to provide early interventions, thus hopefully saving more lives.