The reliability of the clinical examination in predicting hemodynamic status in acute febrile illness in a tropical, resource-limited setting.

Introduction: The clinical examination alone is widely considered unreliable when assessing fluid responsiveness in critically ill patients. Little evidence exists on the performance of the clinical examination to predict other hemodynamic derangements or more complex hemodynamic states. Materials and methods: Patients with acute febrile illness were assessed on admission, both clinically and per non-invasive hemodynamic measurement. Correlations between clinical signs and hemodynamics patterns were analyzed, and the predictive capacity of the clinical signs was examined. Results: Seventy-one patients were included; the most common diagnoses were bacterial sepsis, scrub typhus and dengue infection. Correlations between clinical signs and hemodynamic parameters were only statistically significant for Cardiac Index (r=0.75, p-value <0.01), Systemic Vascular Resistance Index (r=0.79, p-value <0.01) and flow time corrected (r=0.44, p-value 0.03). When assessing the predictive accuracy of clinical signs, the model identified only 62% of hemodynamic states correctly, even less if there was more than one hemodynamic abnormality. Discussion: The clinical examination is not reliable to assess a patient's hemodynamic status in acute febrile illness. Fluid responsiveness, cardiodepression and more complex hemodynamic states are particularly easily missed.