Admission Peripheral Edema, Central Venous Pressure, and Survival in Critically Ill Patients

The clinical significance of peripheral edema has not been well described in critical illness.

To assess the clinical significance of peripheral edema detected on physical examination at the time of hospital admission for patients who were treated in an intensive care unit (ICU).

Using a large inception cohort of critically ill patients, we examined the association of peripheral edema, as documented on hospital admission physical examination, with hospital and 1-year survival.

Of 12,778 patients admitted to an ICU at a teaching hospital in Boston, Massachusetts, 2,338 (18%) had peripheral edema. Adjusting for severity of illness and comorbidities, including pulmonary edema, admission peripheral edema was associated with a 26% (95% confidence interval [CI] = 1.11-1.44, P < 0.001) higher risk of hospital mortality. In those patients whose peripheral edema could be graded, trace, 1+, 2+, and 3+ admission peripheral edema was associated with a 2% (95% CI = 0.80-1.31, P = 0.89), 17% (95% CI = 1.00-1.56, P = 0.05), 60% (95% CI = 1.26-2.04, P < 0.001), and 54% (95% CI = 1.04-2.29, P = 0.03) higher adjusted risk of hospital mortality, respectively, compared with patients without edema. The association was consistent across strata of patients with diabetes, congestive heart failure, sepsis, and premorbid diuretic or calcium channel blocker use. In a subset of patients with central venous pressures measurements obtained within 6 hours of ICU admission, the highest central venous pressure quartile (>13 cm H2O) was similarly associated with a 35% (95% CI = 1.05-1.75, P = 0.02) higher adjusted risk of hospital mortality compared with the lowest quartile (≤7 cm H2O).

Peripheral edema, as detected on physical examination at the time of hospital admission, is a poor prognostic indicator in critical illness. Whether peripheral edema simply reflects underlying pathophysiology, or has an independent pathogenic role, will require further interventional studies.