Peripheral Endothelial Function as a Marker of Systemic Vasodilation in End-stage Liver Disease: Results of a Pilot Study

Patients waitlisted for liver transplant with refractory complications of portal hypertension may not be well served by model of end-stage liver disease sodium score (MELD-Na) scores. Portal hypertension is the result of increased hepatic vascular resistance aggravated by the interplay between vasoactive mediators.¹ Bedsides, measurement reflecting peripheral vasodilation may help to identify these cirrhotic patients at higher risk for worse outcomes.

The EndoPAT test is a noninvasive device that captures endothelial function using peripheral arterial tonometry in response to reactive hyperemia and correlates this to an index, the reactive hyperemia index (RHI).² Low RHI measure represents vasoconstriction associated with atherosclerotic disease. A higher than normal RHI may suggest a vasodilatory state. We conducted a pilot study to compare RHI in patients with end-stage liver disease to normal control population to determine if a vasodilatory state can be measured in cirrhosis. In addition, the RHI was assessed in relation to complications of portal hypertension.

This was a prospective study, approved by the Institutional Review Board, of patients with chronic end-stage liver disease listed for primary, single organ, liver transplantation. Controls from a general population dataset were retrospectively identified and matched 2:1 for age (±5 y) and gender.³ Complications of portal hypertension were included if documented in the patient history or by physical or imaging examination findings at the time of the EndoPAT testing. A history of a >3 therapeutic paracentesis or transjugular intrahepatic portosystemic shunt for an indication of refractory ascites or bleeding was also determined. In a fasting state using EndoPAT 2000 (Itamar Medical Inc., Israel), a finger probe placed on the middle finger with a blood pressure cuff on the proximal arm inflated to 60 mm Hg above baseline systolic pressure for 5 minutes. On cuff deflation, reactive hyperemia occurs resulting in vasodilatation. The RHI, a ratio of signal after cuff deflation with baseline was calculated through computer algorithms. Associations of variables with RHI in its continuous form were done using univariate linear regression models using a 2-tailed t-test. Comparison of the sample with matched controls was done using univariate conditional logistic regression models.

One hundred fifteen patients with mean age 54.8 (±10.4) and 59.1% male gender were matched 2:1 to general population patients. In cirrhotic patients, ascites was noted in 91 (79%) and varices in 93 (81%, 33 with bleeding history). Of 91, 15 patients with ascites had a history of spontaneous bacterial peritonitis (SBP). Mean RHI is demonstrated in Figure 1. RHI correlated with ascites (mean RHI 0.36 higher for those with ascites, 95% CI 0.07-0.64, P = 0.014). RHI >2 was associated with SBP (odds ratio, 3.55; 95% confidence interval [CI], 1.05-16.13; P = 0.041). There was no association with varices or variceal bleeding (P = 0.633 and P = 0.446, respectively). Sixty-seven patients underwent successful liver transplantation, and 20 patients died without transplant. Ascites complicated 19/20 (95%) patients who died, with 2/20 (10%) experiencing significant worsening of their ascites within 30 days of death. Three of the patients who died had a transjugular intrahepatic portosystemic shunt for refractory ascites before the death with varying control of ascites. Univariate Cox regression model did not show a statistically increased risk of death related to ascites (hazard ratio, 2.66; 95% CI, 0.61-11.54; P = 0.190) nor RHI (hazard ratio, 1.14; 95% CI, 0.60-2.16; P = 0.692). The cause of death for patients who died on the waiting list was varied and included sepsis and/or end-stage liver disease (9), cardiopulmonary (3), trauma/bleeding (1), hypoglycemia (1), and complications of malignancy (2). The cause of death was unknown for 4 patients.

FIGURE 1.

Median RHI for the general population compared to patients with or without ascites and with or without SBP. RHI, reactive hyperemia index; SBP, spontaneous bacterial peritonitis.

This pilot study demonstrated a potential relationship between higher RHI values and ascites/SBP. Patients with refractory ascites are often more ill than the MELD-Na score reflects. This data opens the door to further studies evaluating RHI as an additional bedside measure possibly reflecting portal hypertension complications beyond the MELD-Na score. Further larger studies into combinations of MELD–Na with RHI to reflect mortality risk are warranted.