Clinical Radiology (1997) 52, 220-223

Value of the Measurement of Portal Flow Velocity in the Differential

Diagnosis of Asymptomatic Splenomegaly
E. TINCANI, G. CIONI, P. D'ALIMONTE*, A. CRISTANI, F. TURRINI, C. SARDINI, R. ROMAGNOLI* and
E. VENTURA
Departments of Internal Medicine and *Radiology, University of Modena, Italy

Aim and Methods: This prospective study was carried on 20 patients (10 with liver cirrhosis
and 10 with myelo-lymphoproliferative disorders), consecutively admitted to our ward for
splenomegaly and thrombocytopenia, with the aim of evaluating the ability of Duplex-
Doppler ultrasonography (DDUS) to discriminate between congestive splenomegaly and
enlarged spleen caused by haematological disorders.
Results: Comparing the clinical/laboratory and DDUS findings for the two groups, it
emerged that maximum-portal flow velocity (PFV) values revealed the most statistically
significant differences: 17.31 SD 2.48 vs. 28.27 SD 3.53 (cm/s, P < 0.001). Discriminant
analysis showed that max-PFV is the variable which by itself maximizes the separation
between the two groups (F = 71.56; P < 0.0001). The patients with congestive splenomegaly
exhibited lower max-PFV than the controls (17.31 SD 2.48 vs. 26.29 SD 2.38cm/s,
P < 0.001), unlike those with haematological diseases, whose max-PFV values were greater,
albeit not significantly so (28.27 SD 3.53 vs. 26.29 SD 2.38 cm/s, P = 0.161).
Conclusions: DDUS assessment of portal haemodynamics thus proved useful in the
differential diagnosis of splenomegaly in asymptomatic patients since it distinguishes
rapidly and non-invasively between congestive and haemotoiogical splenomegaly. A
lower-than-normal max-PFV value indicates congestive splenomegaly; a highish value,
on the other hand, suggests a splenomegaly of haematological origin. Tincani, E., Cioni, G.,
D'Alimonte, P., Cristani, A., Turrini, F., Sardini, C., Romagnoli, R. & Ventura, E. (1997).
Clinical Radiology 52, 220-223. Value of the Measurement of Portal Flow Velocity in the
Differential Diagnosis of Asymptomatic Splenomegaly

Accepted for Publication 12 July 1996

Splenomegaly is an acknowledged indirect sign of portal
hypertension in liver cirrhosis and a helpful clue to its
diagnosis [1]. As a rule, splenomegaly is associated with
hypersplenism, of which leukopenia, thrombocytopenia and
anaemia are common clinical features [2].
Many diseases in which portal hypertension is absent are
characterized by splenomegaly and hypersplenism, thus
presenting a diagnostic problem [3,4].
Recently [51, portal haemodynamics in patients with
lymphoproliferative and myeloproliferative disorders were
evaluated, but no useful pointers for diagnostic purposes
emerged.
This prospective study was carried out with the aim of
evaluating the ability of Duplex-Doppler ultrasonography
(DDUS) to discriminate between congestive splenomegaly
and enlarged spleen caused by haematological disorders.
PATIENTS AND METHODS
This prospective study was conducted on 20 patients
admitted to our ward for splenomegaly and thrombocyto-
penia between January and December, 1995.
Ten patients (6 male, 4 female) with a mean age of 53.30
SD 8.14 years had liver cirrhosis, while the other 10 (seven
male, three female), with a mean age of 62.80 SD 13.35
Correspondence to: Dr Giorgio Cioni, Department of Internal Medicine,
Policlinico, Largo del Pozzo 71, 41100 Modena, Italy.
© 1997 The Royal College of Radiologists.
years, had lymphoproliferative or myeloproliferative dis-
orders (myelofibrosis of agnogenic myeloid metaplasia,
two; non-Hodgkin lymphoma, seven; Hodgkin disease,
one). The diagnosis was histologically proved in all patients
by liver biopsy, marrow biopsy or pathological examina-
tion of the spleen after splenectomy. The cause of liver
cirrhosis was alcohol in one patient, hepatitis in eight
patients (HBV-related, four; HCV-related, three; HBV-
and HCV-related, one) and mixed (alcohol) in one patient.
According to the Child-Pugh classification for the severity
of liver failure [6], all the patients with liver cirrhosis were
in Child group A.
For each patient the following tests were performed:
haemoglobin, white blood cells, platelets, prothrombin
activity, total bilirubin, albumin, aspartatate aminotransfer-
ase (AST), alanine aminotransferase (ALT), lactate dehydro-
genase (LDH), alkaline phosphatase (AP), and g-glutamyl
transferase (GGT).
All patients underwent: (1) an oesophageal-gastro-
duodenoscopy (OGDS) performed by the same operator,
who recorded the presence or absence of oesophageal
varices and, when present, their grade according to
Dagradi's classification [7]; (2) an echography of the
spleen for the measurement of its bipolar diameter, the
upper normal limit being set at 120mm; (3) a DDUS
study of the portal haemodynamics. The DDUS was per-
formed with a combined ultrasonic system consisting of a
3.5 MHz mechanical sector scanner (ALOKA SSD-650) and
a pulsed Doppler device with a pulse-repetition frequency of
 VALUE OF THE MEASUREMENT OF PORTAL PLOW VELOCITY 221

Table 1 - Characteristics of the patients studied

Congestive Haematological P
splenomegaly disorders

Laboratory findings
Haemoglobin (g/dl) 12.74 SD 1.11 10.60 SD 2.12 0.014
ALT (Lift) 75.70 SD 31.92 39.30 SD 23.27 0.010
LDH (U/I) 380.60 SD 48.13 588.67 SD 233.60 0.013
Duplex-Doppler and ultrasonographic findings
Spleen bipolar diameter (cm) 16.6 SD 1.43 19.7 SD 2.83 0.006
Max-PFV (cm/s) 1 17.31 SD 2.48 28.27 SD 3.53 <0.001
Mean-PFV (cm/s)2 15.30 SD 1.83 22.25 SD 3.03 <0.001
Portal vein diameter (cm)3 1.44 SD 0.25 1.46 SD 0.26 0.86
Portal blood flow (ml/min)4 1531 SD 92 2307 SD 60 0.029
Congestion index (cm/s)5 0.15 SD 0.053 0.096 SD 0.003 0.012

1 Max-PFV (cm/sec): control group 26.29 SD 2.38 vs. congestive splenomegaly P < 0.001; vs. haematological diseases P = 0.161.
2 Mean-PFV (cm/sec): control group 21.61 SD 2.54 vs. congestive splenomegaly P < 0.001; vs. haematological diseases P < 0.615.
3 Portal vein diameter (cm): control group 0.99 SD 0.074 vs. congestive splenomegaly P < 0.001; vs. haematological diseases P < 0.001.
4 Portal blood flow (ml/min): control group 1011 SD 36 vs. congestive splenomegaly P = 0.011; vs. haematological diseases P < 0.001.
5 Congestion index (cm/sec): control group 0.047 SD 0.006 vs. congestive splenomegaly P < 0.001; vs. haematological diseases P < 0.001.

2.1KHz and a filter of 100 Hz (ALOKA UGR-650). All
examinations were performed with the Doppler probe at the
same frequency. The intercostal approach was preferred in
order to minimize the angle of incidence, which was always
less than 60 ° . The study was performed in all cases by the
same examiner following a strictly standardized method [8],
in order to minimize systematic and random error: the
examiner was unaware of the clinical condition of the
patients; measurements were obtained in fasting patients
holding their breath during normal respiration, in order to
avoid digestive and respiratory influences on portal flow.
The portal trunk was scanned longitudinally; a 2 m m
sampling marker was shifted to the centre of the lumen,
1 - 2 c m before the bifurcation of the right and the left
portal branches, thereby avoiding any possible influence
of proximal collateral circuli on PFV. Portal vein diameter
(PV), presence of portal flow, its direction, mean and
maximum PFV were assessed. PFV values were angle-
corrected. Portal blood flow (PBF) was calculated as
the product of portal vein area and mean PFV, while the
congestion index (CI) of the portal vein was calculated
according to Moriyasu [9]. No patient was being treated
with any drug able to influence PBF.
Patients were excluded from the study if they had super-
ficial lymphadenopathy or echographic abnormalities
clearly suggesting a diagnosis of liver cirrhosis or haema-
tological disorder, as were those with liver cirrhosis if: (1)
they had clinical evidence of decompensated liver disease
(ascites, hepatic encephalopathy, total bilirubin > 3.0 mg/dl);
(2) they had oesophageal varices at OGDS; (3) they had
ultrasonographic findings of hepatofugal portal flow, portal
thrombosis or hepatocellular carcinoma.
A DDUS study was also performed on 10 normal volun-
teers used as a control group. All the subjects were informed
of the procedures and agreed to participate in the study.
Continuous data are given as mean and standard deviation,
categorical data as frequency. Student's t-test for unpaired
data, stepwise discriminant analysis and chi-square were
used for statistical analysis. The null hypothesis was
rejected at P<0.05. Statistical analysis was performed
with the SPSS software program.
RESULTS
The laboratory and DDUS parameters that resulted sig-
nificantly different between the two groups of patients are
reported in Table 1. The values of haemoglobin, white blood
cells, platelets, prothrombin activity, bilirubin, albumin,
AST, AP and GGT were not significantly different.
Laboratory test and DDUS findings that showed signifi-
cant differences between the two groups were inserted into
a model of stepwise discriminant analysis so as to identify
the combination of parameters with the greatest predictive
power. At step 0, the variable with the highest F-value was
max-PFV (F = 71.56; P < 0.0001). At step 1, after max-PFV

Table 2 - Discriminant analysis

Variables not in the analysis after step 0 Variables not in the analysis after step 1

Variable F to enter* Variable F to enter*

Haemoglobin 6.43 Haemoglobin 0.0002

ALT 7.77 ALT 3.44
LDH 7.62 LDH 2.52
Spleen bipolar diameter 7.67 Spleen bipolar diameter 0.06
Max-PFV 71.56 Max-PFV *
Mean-PFV 49.89 Mean-PFV 1.21
Portal blood flow 4.48 Portal blood flow 0.17
Congestion index 7.51 Congestion index 0.21

* Minimum F to enter = 3.84; *P < 0.0001.

© 1997 The Royal College of Radiologists, Clinical Radiology, 52, 220-223.

222
40
P < 0.001 P = 0.16
I 1 I I
o°o
"o-- ee
E oo • logical diseases had higher max-PFV values, although not
> 20
o m 41o
x in patients with haematological diseases than in those with
[ P < 0.001
Liver Control Haematological
cirrhosis group diseases
Fig. 1 - Values of maximum Portal Flow Velocity (PFV) of patients with
congestive splenomegaly and of haematological diseases with respect to the
control group of normal volunteers.
had been introduced into the analysis, none of the remaining
values had a greater-than-threshold F-value (Table 2).
Figure 1 illustrates the behaviour of the max-PFV values
of patients with haematological disorders and those with
liver cirrhosis in comparison with that of healthy volunteers.
DISCUSSION
Past [3] and recent [4] diagnostic approaches to spleno-
megaly insist on the importance of the identification of liver
cirrhosis, as it is the most common cause of congestive
splenomegaly [10] and probably the most common cause of
splenomegaly in general, in particular in the context of a
chronic illness [3]. Usually, liver cirrhosis is detected early
by clinical and laboratory studies. In other cases the detec-
tion of oesophageal varices by endoscopy may be of use;
however, spleen size correlates poorly with portal venous
pressure and with the size of oesophageal varices, and
oesophageal varices may be absent in cirrhosis [11]. Inva-
sive procedures such as the measurement of hepatic wedge
pressure [12] and liver biopsy may be needed for diagnosis.
Other possible causes of splenomegaly in asymptomatic
patients are agnogenic myeloid metaplasia, hairy-cell leu-
kaemia, polycythaemia vera, chronic myelocytic leukaemia
and splenic lymphoma, especially in their early stages. In
some cases, the complete blood count and peripheral blood
smear may not immediately suggest the diagnosis.
DDUS has been proposed as a valuable method for the
non-invasive study of portal haemodynamics [13]. The
measurement of PFV is considered to be the most repro-
ducible Doppler-indicator [14]; PFV is correlated to portal
vein pressure in patients with liver cirrhosis [9] and it is a
reliable predictor of liver cirrhosis in patients with suspected
chronic compensated liver disease [15].
Dubois et al. [5] studied portal haemodynamics in lym-
phoproliferative and myeloproliferative disorders, and found a
significantly increased portal venous blood flow in patients
with haematological diseases compared with normal con-
trols, but found no correlation between portal venous blood
flow and hepatic venous pressure gradient, in the absence of
abnormal liver histology.
The present study confirms that in patients with haema-
tological illnesses there is an increase in PBF with respect to
the control group of healthy subjects. Further, discriminant
CLINICAL RADIOLOGY
analysis establishes that DDUS parameters, notably max-PFV,
serve best to differentiate between congestive splenomegaly
• and splenomegaly of haematological origin. Compared with
healthy controls, the patients with congestive splenomegaly
. had lower max-PFV values (17.31SD2.48 vs. 26.29SD
2.38cm/s, P<0.001), while the patients with haemato-
°
significantly so (28.27 SD 3.53 vs. 26.29 SD 2.38 cm/s,
P = 0 . 1 6 1 ) . The PV-diameter was not significantly greater
congestive splenomegaly (1.46 SD 0.26 vs. 1.44 SD 0.25 cm,
I P = 0.86). This is an important result in that the increase in
PV-diameter, hitherto considered to be a pointer to the
presence of liver cirrhosis [16], is non-specific and hence
of no use in the diagnostic assessment of a splenomegaly.
Both patients with cirrhosis [17] and those with lympho-
proliferative and myeloproliferative disorders have increased
portal system blood flow [18]; however, in patients with
liver cirrhosis, the presence of an irreversible intrahepatic
resistance to flow is associated with a reduction in flow in
the portal vein itself.
This study was performed on a small number of patients,
both because of its prospective design and because the strict
selection of cases caused the exclusion of a large number of
subjects. Nevertheless, we think that the results obtained
still have clinical value. In fact, the mean values of PFV
observed in the control and patient groups were similar to
those previously found in larger groups of patients both in
our experience [15] and in that of others [5,9,14].
To conclude, DDUS assessment of portal haemo-
dynamics is seen to be useful in the differential diagnosis
of splenomegaly in asymptomatic patients. Its use is recom-
mended, since it distinguishes rapidly and non-invasively
between congestive splenomegaly and secondary spleno-
megaly due to haematological disorders, making for eco-
nomical use of resources and less discomfort for the patient.
A lower-than-normal max-PFV value indicates conges-
tive splenomegaly; a highish value, on the other hand,
suggests a splenomegaly of haematological origin.
REFERENCES
1 Westaby S, Wilkinson SP, Warren R et aL Spleen size and portal
hypertension. Digestion 1978;17:63-67.
2 Sherlock S, Dooley J. The haematology of liver disease. In: Sherlock S,
Dooley J, eds. Diseases of the liver and biliary system. 9th ed. Oxford:
Blackwell Scientific Publications, 1993:44-61.
3 Eichner E, Whitfield CL. Splenomegaly. An algorithmic to diag-
nosis. Journal of the American Medical Association 1981;246:
2858-2861.
4 Bribre PJ. Spl6nom6galie. Orientation diagnostique. La Revue du
Praticien 1994;44:2069-2077.
5 Dubois A, Dauzat M, Pignodel C et al. Portal hypertension in lympho-
proliferative and myeloproliferative disorders: hemodynamic and his-
tological correlations. Hepatology 1993; 17:246-250.
6 Pugh RNH, Murray-Lyon IM, Dawson LJ et al. Transection of the
esophagus for bleeding esophageal varices. British Journal of Surgery
1973;60:646-649.
7 Dagradi AE, Mehler R, Tan DTD et al. Sources of upper gastro-
intestinal bleeding in patients with liver cirrhosis and large oesopha-
geal varices. American Journal of Gastroenterolology 1970;50:
460-463.
8 D'Alimonte P, Cioni G, Cristani A et al. Duplex-Doppler ultrasono-
graphy in the assessment of portal hypertension. Utility of the measure-
ment of maximum portal flow velocity. European Journal of Radiology
1993;17:126-129.
9 Moriyasu F, Nishida O, Ban N e t aL 'Congestion index' of the portal
vein. American Journal of Roentgenology 1986;146:735-741.
10 Pagliaro L, D'Amico G, Luca A et al. Portal hypertension: diagnosis
and treatment. Journal of Hepatology 1995;23(suppl. 1):36-44.
© 1997 The Royal College of Radiologists, Clinical Radiology, 52, 220-223.
 VALUE OF THE MEASUREMENT OF PORTAL FLOW VELOCITY
11 Amarapurkar S, Parikh SS, Shankaran K et al. Correlation between
splenomegaly and oesophageal varices in patients with liver cirrhosis.
Endoscopy 1994;26:563.
12 Reynolds TB. Measurement of portal pressure and its clinical
application. American Journal of Medicine 1970;49:649-658.
13 Ozaki CF, Anderson JC, Lieberman RP et al. Duplex ultrasonography
as a non-invasive technique for assessing portal hemodynamics.
American Journal of Surgery 1988;155:70-75.
14 Bolondi L, Li Bassi S, Gaiani S e t al. Doppler fiowmetry in portal
hypertension. Journal of Gastroenterology and Hepatology 1990;5:
459-467.
© 1997 The Royal College of Radiologists, ClinicalRadiology, 52, 220-223.
223
15 Cioni G, D'Alimonte P, Cristani A et al. Duplex-Doppler assessment of
cirrhosis in patients with chronic compensated liver disease. Journal of
Gastroenterology and Hepatology 1992;7:382-384.
16 Bolondi L, Gandolfi L, Arienti V e t al. Ultrasonography in the diagnosis
of portal hypertension: diminished response of portal vessels to
respiration. Radiology 1982;142:167-172.
17 Sherlock S. Vasodilatation associated with hepatocellular disease:
relation to functional organ failure. Gut 1990;31:365-370.
18 Shaldon S, Sherlock S. Portal Hypertension in the myeloproliferative
syndrome and the reticuloses. American Journal of Medicine 1962;
32:758-763.