Accuracy of point of care ultrasound to identify the source of infection in septic patients: a prospective study

Abstract

Sepsis is a rapidly evolving disease with a high mortality rate. The early identification of sepsis and the implementation of early evidence-based therapies have been recognized to improve outcome and decrease sepsis-related mortality. The aim of this study was to compare the accuracy of the standard diagnostic work-up of septic patients with an integrated approach using early point of care ultrasound (POCUS) to identify the source of infection and to speed up the time to diagnosis. We enrolled a consecutive sample of adult patients admitted to the ED who met the Surviving Sepsis Campaign (SSC) criteria for sepsis. For every patient, the emergency physician was asked to identify the septic source after the initial clinical assessment and after POCUS. Patients were then addressed to the standard predefined work-up. The impression at the initial clinical assessment and POCUS-implemented diagnosis was compared with the final diagnosis of the septic source, determined by independent review of the entire medical record after discharge. Two hundred consecutive patients entered the study. A final diagnosis of the septic source was obtained in 178 out of 200 patients (89 %). POCUS-implemented diagnosis had a sensitivity of 73 % (95 % CI 66–79 %), a specificity of 95 % (95 % CI 77–99 %), and an accuracy of 75 %. Clinical impression after the initial clinical assessment (T0) had a sensitivity of 48 % (CI 95 % 41–55 %) and a specificity of 86 % (CI 95 % 66–95 %). POCUS improved the sensitivity of the initial clinical impression by 25 %. POCUS-implemented diagnoses were always obtained within 10 min. Instead the septic source was identified within 1 h in only 21.9 % and within 3 h in 52.8 % with a standard work-up. POCUS-implemented diagnosis is an effective and reliable tool for the identification of septic source, and it is superior to the initial clinical evaluation alone. It is likely that a wider use of POCUS in an emergency setting will allow a faster diagnosis of the septic source, leading to more appropriate and prompt antimicrobial therapy and source control strategies.

Appendices

Appendix 1. Informed consent to participate

Appendix 2. Surviving Sepsis Campaign criteria for sepsis

Criteri diagnostici per sepsi. Adapt. Ref. [27].

Appendix 3. Technical approach of POCUS

Descriptive POCUS findings

Anatomic district	Normal findings	Pathological findings
Lung Anterolateral and posterior scans (two anterior, two lateral, one posterior) Convex 3.5–5 MHz probe/linear 5–7 mHz probe
Sub-pleural lung consolidation, presenting a tissutal pattern with dynamic air or multiple hyper-echogenic spots	No	Yes
Focal interstitial syndrome	No	Yes
Presence of pleural fluid	No	Yes
Heart Parasternal view (long and short axis), apical view, subcostal view (4-chambers) Sector 2–2.5 MHz probe
Presence of vegetation on the valve surface	No	Yes
Abdomen
(1) Gallbladder and biliary duct Convex 3.5–5 MHz probe
Wall thickness >4 mm	No	Yes
Pericholecystic fluid	No	Yes
Gallstones/sludge	No	Yes
Echographic murphy sign	No	Yes
Common bile duct >5 mm	No	Yes
(2) Liver Convex 3.5–5 MHz probe
Hepatic abscess
Ascites (primary PBS)	No	Yes
(3) Diverticula Convex 3.5–5 MHz probe/linear 5–7 mHz probe
Presence of diverticula	No	Yes
Wall thickness >3 mm	No	Yes
Inflammatory peri-colonic fat	No	Yes
Presence of abscesses	No	Yes
Peri-colonic free fluid	No	Yes
(4) Appendix Convex 3.5–5 MHz probe/linear 5–7 mHz probe
Total diameter on cross section >6 mm or Wall thickness >3 mm
Non compressible-appendix	No	Yes
Inflammatory peri-appendiceal fat	No	Yes
Presence of abscesses	No	Yes
Peri-appendiceal free fluid	No	Yes
(5) Abdominal-muscle abscesses Convex 3.5–5 MHz probe
Presence of abscesses	No	Yes
(6) Kidney Convex 3.5–5 MHz probe
Hydronephrosis	No	Yes
Presence of renal abscess	No	Yes
Urethorolithyasis	No	Yes
Joints Linear 5–7 mHz probe
Intra-articular fluid	No	Yes