PURPOSE: To test the peripheral arterial vasodilation hypothesis of so
dium retention in cirrhosis. This states that sodium retention is trig
gered by arterial underfilling and predicts that development of sodium
retention will be associated with significant and related declines in
indices of arterial filling that reverse when sodium retention resolv
es. DESIGN: Longitudinal evaluation of a cohort of patients with alcoh
olic liver disease. PATIENTS AND METHODS: Eighteen men, 8 of whom were
studied twice, 3 three times, 2 four times, and 5 rive times (40 betw
een-study comparisons). Between 23 studies, the patients were ascites-
free (Group NN). Ascites spontaneously disappeared between seven studi
es (Group YN), appeared between six studies (Group NY), and remained p
resent between four studies (Group YY). Between-study changes in blood
volume, arterial blood pressure, cardiac output, systemic vascular re
sistance, left atrial volume, left ventricular diastolic diameter, aor
tic root diameter, aortic blood velocity, plasma norepinephrine and at
rial natriuretic factor concentrations, plasma renin activity, and uri
nary sodium excretion were evaluated by paired t-tests. These changes
were also compared among groups by analysis of variance. In addition,
correlations among the changes were sought. RESULTS: Systolic, diastol
ic, and mean arterial pressures, left ventricular diastolic diameter,
aortic root diameter, stroke volume, cardiac output, plasma norepineph
rine concentration, and systemic vascular resistance were unchanged be
tween studies. Left atrial volume increased between studies in Group N
Y. Pulse pressure fell more in Group NY than in Groups NN and YN, prin
cipally as a result of a decline in systolic blood pressure. Plasma no
repinephrine concentration, plasma renin activity, and blood volume ro
se more in Group NY than in Groups NN, YN, and YY. Changes in both sys
tolic and pulse pressures were directly correlated with the change in
sodium excretion but unrelated to the change in plasma norepinephrine
concentration. Changes in plasma norepinephrine concentration and plas
ma renin activity were unrelated to changes in blood pressure, systemi
c vascular resistance, and urinary sodium excretion. CONCLUSIONS: None
of the indices of arterial filling tested except pulse pressure were
related to sodium retention. Reduced pulse pressure is inconsistent wi
th arterial underfilling, as peripheral vasodilation instead increases
pulse pressure by increasing diastolic run-off. These data do not sup
port the hypothesis that arterial underfilling is the stimulus for sod
ium retention in alcoholic cirrhosis.