Role of Perfusion Index as a Predictor of Vasopressor Requirement in Patients With Sepsis

Background Sepsis remains a leading cause of intensive care unit (ICU) admission and is frequently associated with organ dysfunction and high mortality due to impaired microvascular perfusion. As peripheral hypoperfusion is central to septic shock, reliable bedside tools capable of detecting early microcirculatory failure are critically important. Although mean arterial pressure (MAP) and serum lactate are routinely used in sepsis assessment, both have notable limitations. MAP may remain preserved through compensatory vasoconstriction despite ongoing microcirculatory collapse, while lactate elevation is often delayed and influenced by multiple metabolic factors. In this context, the perfusion index (PI), derived from pulse oximetry, has emerged as a direct and quantitative marker of peripheral perfusion. PI represents the ratio of pulsatile arterial blood flow to non-pulsatile tissue blood flow and reflects arterial inflow and microvascular tone. In septic shock, circulatory and microcirculatory derangements, including vasodilation and flow maldistribution, reduce effective arterial pulsatility, leading to a decline in PI. This physiological basis supports the role of PI as an early, non-invasive predictor of vasopressor requirement in patients with sepsis. Methods This prospective observational study was carried out in the ICU of Dhaka Medical College Hospital, enrolling 96 patients with sepsis who met specific inclusion and exclusion criteria. Detailed demographic, clinical, and investigative data were systematically recorded, and both PI and lactate levels were measured before fluid resuscitation. The need for vasopressor support within 24 hours was then assessed and correlated with baseline PI to determine its predictive significance. Result The study population was predominantly middle-aged, with a mean age of 48.2 years, and two-thirds were male, reflecting a typical mixed-ICU demographic. Comorbidities were common, particularly diabetes and combined hypertension-diabetes, alongside notable burdens of coronary artery disease and chronic airway disease. Patients were admitted with a wide range of critical illnesses, spanning trauma, respiratory infections such as pneumonia, neurological emergencies, postoperative states, and metabolic derangements. Vital signs on admission showed physiological stress, with elevated heart and respiratory rates and a broad range of arterial pressures, nearly half presenting with MAP <65 mmHg. When comparing vasopressor with non-vasopressor groups, heart rate (p=0.001) and MAP (p<0.001) showed significant differences, identifying them as strong markers of hemodynamic decline. PI values were split between low and high groups, but most hypotensive patients clustered within the low-PI category, illustrating a clear physiological relationship. The PI ≤0.3 threshold identified nearly all patients with MAP <65 mmHg, reinforcing its validity as an early perfusion marker. Diagnostic performance metrics further demonstrated PI's strength, achieving a sensitivity of 91.3%, specificity of 94%, and high predictive values in both directions. Conclusion PI serves as a reliable early predictor of vasopressor requirement in patients with sepsis before fluid resuscitation.