Ultrasound Assessment of the Hepatic Vasculature

Hepatic veins

The Doppler signal from the left hepatic vein is optimized with the transducer in a subxiphoid approach. The right hepatic vein is optimized from a lateral intercostal approach near the mid axillary line. Optimization of the middle hepatic vein signal is more variable, ranging from subxiphoid to intercostal near the anterior axillary line, to subcostal. In all of these locations, the transducer is generally positioned so that forward flow in the hepatic veins (i.e., out of the liver and toward the heart) is away from the probe and is displayed below the baseline, while reversed flow (away from the heart and into the liver) is above the baseline.

With that in mind, it is easiest to understand the morphology of the hepatic vein waveform by relating it to activity in the right atrium. When the right atrium contracts, forward flow from the hepatic veins into the inferior vena cava (IVC) and right atrium slows and eventually reverses. The reversal of flow produces a short period of flow above the baseline. As the right atrium relaxes, there is relatively rapidly accelerating flow from the liver into the atrium. This is reflected in the waveform as a rapid down slope below the baseline. With progressive atrial filling, the velocity of flow from the hepatic veins into the atrium starts to slow and the Doppler signal starts to approach the baseline. This deceleration of hepatic vein outflow continues until the tricuspid valve opens. At this point there is a short period of passive atrial emptying into the ventricle, which produces a second phase of accelerating hepatic vein flow into the atrium and another down slope below the baseline. The right atrium then starts to contract, and the hepatic vein flow toward the atrium slows and eventually reverses ( Fig. 27.1 ).

It is important to recognize that deep inspiration can blunt the normal hepatic vein pulsatility and should be avoided when possible. To improve demonstration of the normal waveform morphology, it is best to obtain the waveform at the end of expiration. Specific instructions to the patient that tend to work well are “take a normal breath in, take a normal breath out, stop breathing,” and then obtain the waveform.

Predictably, right heart failure and alterations in right atrial filling and emptying cause changes in the hepatic vein waveform. In patients with right heart dysfunction, hepatic vein pulsatility increases and the retrograde pulses are exaggerated, producing a “W” shaped waveform. With tricuspid regurgitation, right ventricular contraction during systole produces retrograde flow from the ventricle into the right atrium and from the atrium into the hepatic veins. This causes an inverted systolic peak in the hepatic vein waveform ( Fig. 27.2A ).

Portal vein

The normal portal vein demonstrates continuous antegrade flow and provides approximately 75% of the blood supply to the liver. Minor degrees of respiratory phasicity can be seen, although they may be difficult to appreciate with Doppler techniques when sampling is performed during suspended respiration.

Because the hepatic sinusoids separate the portal veins from the heart, the degree of portal vein pulsatility related to cardiac activity is considerably less than the hepatic veins. However, some degree of portal vein pulsatility is normal and is well displayed on portal vein waveforms ( Fig. 27.3 ). Pulsatility can be quantified by using an index called the venous pulsatility index (VPI). The VPI is analogous to the arterial resistive index and is calculated as the difference between the maximum velocity and the minimum divided by the maximum velocity. A very pulsatile waveform, where the minimum velocity reached the baseline (i.e., 0 cm/s), would have a VPI of one. A completely nonpulsatile portal vein waveform would have a VPI of zero. Gallix et al. showed that the mean VPI was 0.48 (± 0.31) in a group of normal individuals. A VPI of 0.48 means the minimum velocity is approximately half the maximum velocity.

Another way to quantify portal flow pulsatility is a simple ratio between the minimum and peak flow. Using this approach, a flat waveform would have a portal vein pulsatility of one, and a very pulsatile waveform, where the minimum velocity dropped to the baseline, would have a portal vein pulsatility of zero. Wachsberg et al. showed that 64% of normal patients had a portal vein pulsatility of less than 0.54. That means that in a majority of patients, the minimum velocity was less than half the peak velocity. Increased pulsatility in the portal vein is more prominent in thin individuals. As with the hepatic vein pulsatility, portal flow pulsatility can be blunted by a deep inspiration.

Helical flow is occasionally seen in the portal vein. It occurs in 2.2% (3/135) of normal individuals and in 20% (8/41) of patients with chronic liver disease who are being evaluated for liver transplantation. It is also seen following liver transplantation, TIPS, and in the setting of portal vein stenosis. It is important to recognize because it can be confused with portal vein flow reversal ( Fig. 27.4 ). This pitfall can be avoided by determining the direction of blood flow both proximal and distal to the focal area of helical flow.

Right heart failure and tricuspid regurgitation may produce exaggerated portal vein pulsatility. Given the degree of pulsatility that can be seen in normal patients, however, cardiac dysfunction should probably not be considered unless the portal pulsatility is so great that the minimum velocity reaches zero or reverses (i.e., the VPI is at least one) (see Fig. 27.2B ). It is also important to correlate the portal vein waveform with other sonographic signs of heart dysfunction, including enlargement of the hepatic veins and IVC, and alterations in the hepatic vein waveform.

There is considerable variation in the reported value for normal portal vein velocity. This variation is at least partially dependent on whether the maximum velocity or the time-averaged mean velocity is being reported. Patriquin el al. found that the maximum portal vein velocity ranged from 8 to 18 cm/s in normal fasting adults and increased from 50% to 100% after eating. Haag et al. found that the normal maximum portal vein velocity was 26.5 ± 5.5 cm/s. Abu-Yousef et al. and Kok et al. found the maximum portal vein velocity ranged from 16 to 31 cm/s (mean 22 cm/s) and from 11 to 39 cm/s (mean 23 cm/s), respectively. Zironi et al. found the normal mean portal velocity was 19.6 ± 2.6 cm/s. Cioni et al. found the maximum velocity was 26.7 ± 3.2 cm/s and the mean velocity was 22.9 ± 2.8 cm/s. They considered the normal range for maximum portal vein velocity to be from 20 to 33 cm/s. This variation in the reported range of normal makes it difficult to rely on portal velocities as a sign of portal hypertension. Very low velocities are a good indicator of portal hypertension; however, velocities in the normal range do not exclude the diagnosis.

Alterations in the normal hepatic and portal vein hemodynamics in a normal pregnancy include a decrease in the maximum flow velocity and an increase in the incidence of monophasic Doppler waveforms in the hepatic and portal veins.

Hepatic artery

The hepatic artery waveform has a low-resistance profile with broad systolic peaks, gradual deceleration from systole to diastole, and well-maintained diastolic flow throughout the cardiac cycle. This is similar to that of other solid parenchymal abdominal organs (i.e., kidneys and spleen). The normal hepatic artery resistive index ranges from 0.5 to 0.7.

Venous diameter

There are a number of gray-scale ultrasound signs of portal hypertension. Engorgement of the portal vein and its tributaries is an indicator of elevated pressures ( Fig. 27.5 ). Goyal et al. prospectively compared portal vein diameter, measured at its broadest point, in 100 healthy subjects and 50 patients with portal hypertension. Keeping physiologic variables known to affect portal vein flow (such as fasting state, supine position, and deep inspiration) similar in both groups, they found that the upper limit of normal for portal vein diameter was 16 mm. Using this cutoff value, they achieved an overall sensitivity of 72%, accuracy of 91%, and specificity of 100% in diagnosing patients with suspected portal hypertension. Others have proposed 13 mm as the cutoff for the upper limit of normal portal vein diameter ; however, Stamm et al. showed that the normal mean portal vein diameter measured on computed tomography (CT) at the point of maximum diameter (at least 1 cm distal to the confluence of the splenic and superior mesenteric veins and 1 cm proximal to the first branch of the main portal vein) was 15.5 mm, which was significantly larger than the accepted upper limit of 13 mm. In addition, they found that contrast-enhanced main portal veins are significantly larger (0.56 mm) than unenhanced portal veins. Uncertainty regarding the normal value of portal vein diameter is one of the reasons that this parameter is not relied on to diagnose portal hypertension. Although it is true that an unusually large portal vein is a reliable sign of portal hypertension, it is also true that a normal-sized portal vein does not exclude the diagnosis.

If one assumes that elevated portal pressure maximizes venous distention, it follows that little or no additional distention will occur when the portal vein outflow is indirectly restricted by sustained inspiration. Lack of caliber change of the splenic and mesenteric veins during respiration has also been investigated. In one study, this approach had a sensitivity of 80% and a specificity of 100% in diagnosing portal hypertension. As with portal vein diameter measurements, this method has not gained widespread acceptance, likely because of a combination of interobserver variability and difficulties in measurement accuracy.