sensitivity and specificity

Importance Accurate assessment of intravascular volume facilitates management decisions about fluid management in patients with volume overload.

Objective To identify the most accurate clinical examination, radiographic, and laboratory findings for assessing volume overload in nonintubated patients.

Data Sources and Study Selection MEDLINE was searched (1946 to January 6, 2026) to identify peer-reviewed English-language studies about the diagnostic accuracy of the clinical examination of spontaneously breathing patients with intravascular volume overload.

Data Extraction and Synthesis Three authors independently extracted data for each finding and calculated sensitivity, specificity, and likelihood ratios (LRs). A 2-level mixed logistic regression model was used to pool estimates.

Results Forty studies, involving 11 490 adult patients, were included, with a prevalence of volume overload of 35% to 69%. Thirty-three of those studies evaluated patients with dyspnea. Prevalence of volume overload was more likely when the physical examination revealed jugular venous distention with the highest point of pulsation more than 3 cm in a vertical line above the sternal angle (LR, 4.1 [95% CI, 2.9-5.6]; specificity, 92%), lower extremity edema (LR, 2.2 [95% CI, 1.5-3.1]; specificity, 80%), or crackles on auscultation (LR, 2.7 [95% CI, 1.7-4.5]; specificity, 81%). Vascular congestion on chest radiography increased the likelihood of intravascular volume overload (LR, 5.9 [95% CI, 2.9-12.0]; specificity, 91%). Point-of-care ultrasonography that identified bilateral pulmonary B-lines suggested volume overload (LR, 4.0 [95% CI, 2.6-6.1]; specificity, 77%), and absence of pulmonary B-lines made volume overload unlikely (LR, 0.09 [95% CI, 0.04-0.23]; sensitivity, 93%). Inferior vena cava collapsibility index of less than 50% increased the likelihood of volume overload (LR, 3.9 [95% CI, 2.5-6.1]; specificity, 79%), and a collapsibility index of at least 50% made it less likely (LR, 0.22 [95% CI, 0.11-0.45]; sensitivity, 82%). Point-of-care ultrasonographic measurement of jugular venous pressure (JVP; >8 cm) also increased the likelihood of volume overload (LR, 2.8 [95% CI, 2.2-3.5]; specificity, 71%), although JVP of 8 cm or less identified patients less likely to have volume overload (LR, 0.26 [95% CI, 0.20-0.33]; sensitivity, 81%). A plasma brain-type natriuretic peptide (BNP) level of 100 ng/mL or higher was the single best test to identify those most likely to have volume overload (LR, 6.9 [95% CI, 2.4-20.4]; specificity, 87%), and a normal value made it less likely (LR, 0.14 [95% CI, 0.08-0.24]; sensitivity, 87%).

Conclusions and Relevance A BNP level of 100 ng/mL or higher and presence of vascular congestion on chest radiography may be the most useful tests to identify patients with volume overload. Absence of pulmonary B-lines using point-of-care ultrasonography or BNP levels of less than 100 ng/mL may be most useful to exclude volume overload.