ORIGINAL ARTICLES LIVER, PANCREAS AND BILIARY TRACT

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2008;6:1129 1134 ORIGINAL ARTICLES LIVER, PANCREAS AND BILIARY TRACT Spleen Enlargement on Follow-Up Evaluation: A Noninvasive Predictor of Complications of Portal Hypertension in Cirrhosis ANNALISA BERZIGOTTI, PAOLA ZAPPOLI, DONATELLA MAGALOTTI, CAROLINA TIANI, VALENTINA ROSSI, and MARCO ZOLI Dipartimento di Medicina Interna, Cardioangiologia, Epatologia, Università di Bologna, Bologna, Italy Background & Aims: Splenomegaly is observed in most but not all patients with cirrhosis, and has been detected more often in patients showing complications of portal hypertension. We aimed to test the hypotheses that spleen enlarges over time in cirrhosis, and that a progressive enlargement may be associated with portal hypertension related events. Methods: A total of 127 cirrhotic patients (Child Pugh, 6.7 1.7; range, 5 11), observed at our center and followed-up clinically, endoscopically, and with periodic abdominal ultrasound for at least 1 year, were included. Spleen diameter was recorded at each ultrasound examination. The change of spleen diameter over time was calculated. The occurrence of clinical complications of cirrhosis on follow-up evaluation was recorded. Results: At inclusion, spleen diameter was 14.9 3.1 cm; 83% of the patients had splenomegaly. Spleen was larger in patients with decompensated disease (n 39) versus patients with compensated disease (n 88) (16.1 3.5 vs 14.5 2.7; P.012). The mean follow-up period was 53 37 months. Spleen progressively enlarged over time (analysis of variance, P <.0001). A total of 46.4% of patients showed a spleen enlargement of 1 cm or more at 1 year. Over 5 years of follow-up evaluation patients showing spleen enlargement showed a higher actuarial probability of esophageal varices formation (84.6% vs 16.6%; P.001) and growth (63.3% vs 20.6%; P.001). Among patients with compensated cirrhosis at inclusion, those showing a spleen enlargement had a higher actuarial probability of developing the first clinical decompensation of cirrhosis (51.1% vs 19.5%, P.002). Conclusions: Spleen enlargement at follow-up evaluation outlines a subgroup of cirrhotic patients at higher risk of complications of portal hypertension. Noninvasive monitoring of spleen diameter allows a prognostic stratification of cirrhotic patients. Splenomegaly is observed in 50% to 75% of cirrhotic patients, 1,2 and is caused by congestion as a result of portal hypertension, tissue hyperplasia, and fibrosis. 3 The increase in splenic blood flow that occurs as a consequence of spleen enlargement contributes to the maintenance and worsening of portal hypertension because it actively congests the portal system. Spleen size can be evaluated reliably by ultrasound (US), 4 and splenomegaly is the ultrasonographic finding associated more frequently with portal hypertension. 5 It has been observed that decompensated patients show splenomegaly more often than those with compensated cirrhosis, 6 and in several studies splenomegaly, assessed by physical examination, US, or computerized tomography, has been shown to correlate with the presence of esophageal varices, 7 10 and with their endoscopic grade. 6,11 Moreover, spleen volume measured by magnetic resonance imaging has been shown to independently predict clinical progression of cirrhosis. 12 Given these observations, it can be hypothesized that spleen dimension increases over time in cirrhotic patients as an indirect sign of onset and worsening of portal hypertension, and that an increase in spleen size thus may be a predictor of the formation and growth of esophageal varices, and of decompensation of cirrhosis in compensated cirrhotic patients. This study aimed to test these hypotheses. Materials and Methods Patients A retrospective, single-center cohort study was designed. All patients with histologic or clinically and laboratory proven cirrhosis who were observed and followed up at our unit from January 1987 to January 2004 were considered. Patients were included if they underwent at least 3 consecutive abdominal US controls (first observation, at 6 months, and at 1 year) and at least 1 year of clinical and endoscopic follow-up thereafter. The mean follow-up period was 53 37 months. Exclusion criteria were the presence of portal vein thrombosis, multifocal hepatocellular carcinoma, and lack of clinical and/or endoscopic follow-up evaluation. Overall, 183 patient files from patients at first visit to our hepatology outpatient clinic were analyzed; 36 patients were excluded because of the presence of portal vein thrombosis (n 9), multifocal hepatocarcinoma (n 12), or inadequate endoscopic, ultrasonographic, and/or clin- Abbreviations used in this paper: ANOVA, analysis of variance; CI, confidence interval; HVPG, hepatic venous pressure gradient; MELD, Model for End-Stage Liver Disease; OR, odds ratio; US, ultrasound. 2008 by the AGA Institute 1542-3565/08/$34.00 doi:10.1016/j.cgh.2008.05.004

1130 BERZIGOTTI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6, No. 10 Table 1. Main Baseline Clinical, Biochemical, and Ultrasonographic Characteristics of the Patients Included in the Study Clinical variable Overall (n 127) Splenomegaly (n 105) Normal spleen size (n 22) Age (y) 59 11 57 11 66 8 Sex (male/female) 76/51 67/38 8/14 Etiology (hepatitis C virus/alcohol/hepatitis B virus/other) 61/30/11/25 50/28/8/19 11/2/3/6 Child Pugh score 6.7 1.7 6.8 1.8 6.1 1.3 a Child Pugh class (A/B/C) 65/51/11 52/43/10 13/8/1 MELD score 10 5 10 4 9 6 Bilirubin level (mg/dl) 1.7 1.3 1.8 1.2 1.5 1.9 International normalized ratio 1.33 0.26 1.34 0.27 1.24 0.13 Albumin level (g/dl) 3.4 0.6 3.4 0.5 3.4 0.6 Creatinine level (mg/dl) 0.9 0.1 0.9 0.2 0.8 0.1 Platelets 10(9)/L 92 40 89 39 110 45 b Ascites (%) 26.1 29.5 13.6 a Encephalopathy (%) 3.3 2.8 4.5 Previous decompensation (%) 35.4 39.0 18.1 Esophageal varices (F0/F1/F2/F3) 43/63/13/8 30/54/10/8 13/9/3/0 Gastric varices (%) 2.4 2.8 0 Gastropathy (%) 9.8 11.4 0 Spleen size (cm) 14.9 3.1 15.8 2.5 10.5 1.1 b a P.10 vs patients with splenomegaly. b P.05 vs patients with splenomegaly. ical follow-up evaluation (n 35). A total of 127 cirrhotic patients were finally included. The following events occurring during the follow-up period were recorded from patient files: formation and growth of esophageal varices; decompensation of cirrhosis, defined as portal hypertension-related bleeding, ascites, spontaneous bacterial peritonitis, hepatic encephalopathy, and jaundice; onset of hepatocellular carcinoma; and liver disease related death. The severity of liver disease at inclusion and during follow-up evaluation was assessed by Child Pugh score 13 and by Model for End-Stage Liver Disease (MELD) score. 14 Table 1 shows the main clinical data of the included patients. Thirty-nine patients (31%) had a decompensated liver disease at inclusion, whereas 88 (69%) had a compensated cirrhosis. According to clinical stages of cirrhosis, 15 30 patients (24%) belong to stage 1, 62 (49%) to stage 2, 30 (24%) to stage 3, and 5 (4%) to stage 4. Sixty-five patients belonged to Child Pugh class A, 51 to class B, and 11 to class C. Thirty-nine patients had experienced previous episodes of decompensation of liver disease (ascites in 29, variceal bleeding in 5, and encephalopathy in 5). At the first visit only 1 patient was on -blockers owing to a previous episode of variceal bleeding. Platelet count was 92 40 (10(9)/L; range, 26 217). Seventy percent of patients had a platelet to spleen ratio of less than 909. 16 As for the endoscopic findings, 84 (66%) patients had esophageal varices (small in 63, medium in 13, and large in 8). Red color signs at endoscopy were found in 21 of 84 patients (25%). Gastric varices and portal gastropathy were detected, respectively, in 3 (2.4%) and 12 (9.8%) patients. Methods Ultrasonographic measure of the bipolar longitudinal diameter of the spleen (expressed in cm) is looked at routinely in our center at each examination and photographic documentation of the diameter was available in every case. Spleen size was measured after an overnight fast, using a duplex Doppler ultrasound equipment (AU-5; Esaote, Genoa, Italy, and Technos Esaote, Genoa, Italy) with a 4.5- to 7-MHz convex probe. Spleen was considered of normal size when the diameter was less than 12 cm. 17 Measures of spleen diameter at each examination were recorded in a database and the change between measurements at 1-year intervals was calculated. Intraobserver and interobserver coefficients of variation for US measurements at our laboratory have been evaluated previously, and did not exceed, respectively, 5% and 10%. Endoscopic evaluation was performed by a single experienced operator (M.Z.) at our unit at the follow-up evaluation. Beppu et al 18 classification was used to grade esophageal varices. In this classification esophageal varices are graded as F1 (small, flattened/disappear with insufflation), F2 (tortuous, protrude less than a third of the esophageal lumen), and F3 (protrude more than a third of the esophageal lumen). Variceal growth in the follow-up evaluation was defined as an increase of the occupancy of the lumen, according to the earlier-mentioned classification. According to the clinical practice guidelines of our center, patients received a first endoscopy at diagnosis. Patients with no varices on first endoscopy were followed up endoscopically at 2- to 3-year intervals and patients with small varices were followed up at 1- to 2-year intervals. When medium or large varices were diagnosed, -blockers were started for primary prophylaxis and no other endoscopy was performed unless a bleeding episode was seen. This study was approved by the Senior Staff Committee of the University Hospital, a board that regulates noninterventional studies and is comparable with an institutional review board. Statistical Analysis Statistical analysis was performed using SPSS 14.0 statistical package (SPSS Inc, Chicago, IL). All results are expressed as frequencies or as mean SD. Comparisons between groups were performed by the chi-square test for frequencies and by the Student t test for continuous variables. The Kruskal Wallis test was used when appropriate. Comparisons of spleen dimension of different controls during the follow-up evaluation were performed with the Student t test for paired data. Correlations were made using the Pearson test. Changes of spleen diameter

October 2008 SPLEEN ENLARGEMENT IN CIRRHOSIS 1131 Figure 1. Spleen dimension on US follow-up evaluation in the studied population. As shown, spleen enlarged over time in cirrhotic patients (ANOVA, P.0001). and platelet count during the follow-up period were analyzed with repeated-measure analysis of variance (ANOVA). The cumulative risk of developing the analyzed end points was evaluated by Kaplan Meier curves. A comparison between patients with and without spleen enlargement was made using the log-rank test. Multivariate analyses were performed by using the stepwise binary logistic regression method. A P value of less than.05 was considered statistically significant. As for the change in spleen diameter at 1 year, we arbitrarily chose a cut-off value of 1 cm as a significant increase. This was done because of the ease in remembering this cut-off value, and because smaller increases might be caused by intraobserver or interobserver variability. Results Spleen Dimension at First Observation The longitudinal diameter of the spleen at inclusion was a mean of 14.9 3.1 cm (range, 8.0 27.0 cm). Twenty-two (17%) patients had a normal spleen diameter, whereas 105 (83%) had an enlarged spleen (Table 1). Spleen size was significantly greater in patients with decompensated versus compensated cirrhosis (16.1 3.5 vs 14.5 2.7; P.012). Spleen diameter showed a slight correlation with Child Pugh score (R 0.191; P.03), and with esophageal varices size (Kruskal Wallis test, 5.85; P.05). Patients with red color signs at endoscopy showed a larger spleen than patients without red signs (16.5 4.0 vs 14.8 2.5 cm; P.018). Spleen was slightly larger in patients with ascites versus those without ascites (15.6 3.5 vs 14.7 2.9 cm; P.14), but data did not reach statistical significance. Clinical Course of the Patients in the First Year and Spleen Dimension Over the Follow-Up Period Within the first year of observation patients had an uneventful clinical course; in 5 patients presenting with ascites at baseline the ascites improved. Overall, 29 (22.8%) patients received beta-blockers either for the primary or secondary prophylaxis of variceal bleeding in the year between basal and 1-year measurement of spleen diameter. One patient was already taking beta-blockers on first observation as a secondary prophylaxis after a previous episode of variceal bleeding diagnosed in another hospital. In the remaining 4 patients who had had variceal bleeding beta-blockers were added on the first visit at our center. Similarly, in the 24 patients in whom mediumlarge varices were detected at first endoscopy, beta-blockers were added on the first medical visit thereafter. Spleen dimension increased over the follow-up period (Figure 1; ANOVA, P.0001). At 1 year the spleen mean diameter showed a mean increase of 0.9 0.9 cm versus baseline. Spleen increased in size in 86 (67.7%) patients; in 27 (21.3%) of them the increase was slight (change, 1 cm), whereas in 59 (46.4%) it was more pronounced (change, 1 cm). In 39 (30.7%) patients spleen diameter was unchanged. In 2 (1.6%) patients spleen reduced in size. Clinical and laboratory data at baseline were similar in patients who did and did not develop significant spleen enlargement (ie, 1 cm). In the first year after baseline observation, the course of liver disease was similar in patients who did and did not develop significant spleen enlargement (Table 2). At 1 year, patients with enlarged spleen showed a slightly lower platelet count (P.07) and albumin level (P.08) as compared with patients with stable spleen size, but data did not reach full statistical significance. Patients undergoing -blocker therapy showed a similar rate of spleen enlargement at 1 year as compared with untreated patients. The enlargement of spleen during the follow-up period is shown in Figure 1. The finding of spleen enlargement at 1 year was confirmed in all but 3 patients in the further follow-up period because it was confirmed in all but 3 patients. Spleen Enlargement and Formation, Growth, and Rupture of Esophageal Varices A total of 103 patients had no varices (n 43) or small varices (n 63) at first endoscopy, and received an endoscopic Table 2. Clinical and Biochemical Data of the Studied Patients at 1 Year Spleen enlargement 1 cm(n 68) Spleen enlargement 1 cm(n 59) Child Pugh score 6.7 1.7 6.8 1.6 MELD score 10 5 9 4 Bilirubin level (mg/dl) 1.9 2.0 1.8 1.9 International normalized ratio 1.28 0.22 1.37 0.38 Albumin level (g/dl) 3.5 0.4 3.3 0.6 Creatinine level (mg/dl) 0.98 0.19 0.92 0.32 Platelets 10(9)/L 100 56 78 28 a,b Ascites (%) 20.5 23.7 Encephalopathy (%) 2.9 3.3 Improvement of ascites (n) 3 2 Added -blocker therapy (%) 13.2 a 16.9 a Median follow-up (mo) 36 44 NOTE. At 1 year examination US examination revealed spleen enlargement of 1 or 1 cm. a P.05 vs baseline observation. b P.10 vs spleen enlargement 1 cm.

1132 BERZIGOTTI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6, No. 10 Also in the setting of variceal growth the separate analysis of compensated and decompensated patients showed comparable results. Among patients with small varices and compensated cirrhosis (n 48), 21 developed large varices; among patients with no varices and decompensated cirrhosis (n 15), 8 developed large varices. In both groups the cumulative probability of variceal formation was higher in patients with increased spleen size (for compensated patients, 60% vs 28.5%, P.033; for patients with decompensated disease, 87.5% vs 12.5%, P.002). On a multivariate analysis including Child Pugh score, MELD score, alcohol consumption, ascites, spleen size at baseline, and spleen enlargement at 1 year, only spleen enlargement (OR, 6.90; 95% CI, 2.11 22.57; P.001) was associated independently with the growth of esophageal varices. During the follow-up period, 7 (5%) patients experienced variceal bleeding. There was no significant difference in variceal hemorrhage rate between patients with and without an increase in spleen size (18.3% vs 9.3%; P.213; log-rank test, 0.10, P.746). Figure 2. The actuarial probability of formation of varices at 5 years was higher in patients with increased spleen size at 1 year (grey line)vs those with stable spleen size (black line). Log-rank test, 8.47; P.004. *Months of follow-up evaluation after the clinical, laboratory, and US assessment at 12 months. follow-up evaluation. In the follow-up evaluation, 17 of 43 (39.5%) patients showed variceal formation, 29 of 63 (46.0%) showed variceal growth. Over 5 years of follow-up evaluation, patients showing a spleen enlargement of 1 cm or greater at 1 year had a higher cumulative probability of variceal formation compared with patients with stable spleen size (84.6% vs 16.6%; 2, 11.5; P.001). The actuarial probability of variceal formation was also significantly higher in this group (Kaplan Meier log-rank test, 8.47; P.004; Figure 2). The separate analysis of compensated and decompensated patients showed similar results. Among patients with no varices and compensated cirrhosis (n 30), 12 developed varices; among patients with no varices and decompensated cirrhosis (n 13), 5 developed varices. In both groups the cumulative probability of variceal formation was higher in patients with increased spleen size (compensated disease, 76.9% vs 18.1%; P.004; decompensated disease, 80% vs 0%; P.006). On a multivariate analysis including Child Pugh score, MELD score, alcohol consumption, platelet count, spleen size at baseline, and spleen enlargement at 1 year, only spleen enlargement (odds ratio [OR], 12.00; 95% confidence interval [CI], 1.8 91.08; P.016) was associated independently with the formation of esophageal varices. Variceal growth also occurred more frequently in patients with spleen enlargement than in patients with unchanged spleen size: 63.3% versus 20.6%; 2, 10.9; P.001. The 5-year actuarial probability of variceal growth showed a trend to be greater in this group (Kaplan Meier log-rank test, 3.42; P.06; Figure 3). Spleen Enlargement and First Decompensation of Cirrhosis Among the 88 compensated patients, 30 (34%) developed the first clinical decompensation of cirrhosis over 5 years of follow-up evaluation (22 ascites, 4 variceal hemorrhage, 2 spontaneous bacterial peritonitis, 2 encephalopathy). Patients showing spleen enlargement had a higher rate of first clinical decompensation compared with patients with stable spleen size: 51.1% versus 17.8% (P.002). The 5-year actuarial probability of first decompensation of cirrhosis was greater in patients showing spleen enlargement (Kaplan Meier log-rank test, 5.66; P.017; Figure 4). Patients receiving -blockers showed a similar probability of clinical Figure 3. The actuarial probability of growth of varices at 5 years was higher in patients with increased spleen size at 1 year (grey line) vs those with stable spleen size (black line). Log-rank test, 3.42; P.06. *Months of follow-up evaluation after the clinical, laboratory, and US assessment at 12 months.

October 2008 SPLEEN ENLARGEMENT IN CIRRHOSIS 1133 than 100,000. Conversely, patients with no varices and a platelet count of greater than 100,000 at 1 year showed a lower rate of variceal formation (36% vs 67%; P.14) and variceal growth (14% vs 86%, P.04) as compared with patients with platelet count of less than 100,000. Because of these observations we included spleen size increase and platelet count of more than 100,000 in a logistic regression analysis to evaluate if only one or both maintained an independent predictive value for the occurrence of variceal formation and growth. As for variceal formation the model selected only spleen size increase (OR, 22.50; 95% CI, 2.60 194.50; P.002), whereas both spleen size increase (OR, 9.045; 95% CI, 2.328 35.139; P.012) and platelet count of more than 100,000 at 1 year (OR, 4.19; 95% CI, 0.88 19.94; P.047) were found to be associated independently with variceal growth. Among patients with compensated disease, 35% had a platelet count of more than 100,000 at 1 year. These patients showed a similar figure of first decompensation of cirrhosis as compared with patients with a lower platelet count (51% vs 48%, P.80). Figure 4. The actuarial probability of first clinical decompensation of cirrhosis at 5 years in patients with compensated disease at inclusion was higher in patients with increased spleen size at 1 year (grey line)vs patients with stable spleen size (black line) (log-rank test, 5.66; P.017). *Months of follow-up evaluation after the clinical, laboratory, and US assessment at 12 months. decompensation of the disease on follow-up evaluation as compared with patients who were not on -blockers (14.2% vs 38.4%; 2, 1.19; P.275). On multivariate analysis including the objective components of Child Pugh score, MELD score, alcohol consumption, presence of esophageal varices, spleen size at baseline, and spleen enlargement at 1 year, albumin level (OR, 0.13; 95% CI, 0.03 0.48; P.002) and spleen enlargement (OR, 3.20; 95% CI, 0.94 10.99; P.045) remained independent predictors of clinical decompensation. Nineteen (15%) patients died from liver disease related causes during the follow-up period. Mortality was similar in patients with and without increased spleen size (17% vs 12%, P.497; log-rank test, 4.02, P.526). Correlation Between Spleen Size and Platelet Count Spleen size and platelet count at baseline showed a negative correlation (R 0.289; P.002). Platelet count decreased over the follow-up period in the overall population (ANOVA, P.031). When patients were stratified according to the presence or absence of spleen enlargement at 1 year, platelet count was found to decrease significantly over the follow-up period only in patients with increased spleen size (ANOVA, P.528 for patients with stable spleen size; P.030 for patients with increased spleen size). Forty-two (33%) patients had a platelet count of more than 100,000 at baseline, which persisted at 1 year in 30; in addition, 12 patients with a platelet count of less than 100,000 at baseline showed values of more than 100,000 at 1 year. No differences in the figures of variceal formation and growth were seen between patients with a platelet count at baseline of greater than and less Discussion In the present study we report 2 novel findings: first, spleen enlarges over time in patients with cirrhosis, and, second, the observation of a spleen enlargement on US follow-up evaluation is associated with the appearance and growth of esophageal varices, and with the occurrence of a first decompensation of cirrhosis in patients with compensated disease. Spleen enlargement over time in cirrhosis is not an unexpected finding because several studies have pointed out that the number of patients presenting with splenomegaly increases in patients with decompensated disease, who also have a longer duration of liver disease. 6 10 Spleen size correlated with the severity of cirrhosis and with the presence and grade of esophageal varices also in the present series, confirming these previous observations. Splenomegaly, defined as a longitudinal diameter greater than 12 cm, 17 was very frequent in our population. We observed splenomegaly in 84% of the patients, which exceeds previously reported prevalences. This may be owing to a different cut-off rate for the definition of splenomegaly, which in other studies has been set variably at 13 to 14 cm. 2,19 As for the prognostic value of the increase of spleen size, in the statistical analysis of the present study we arbitrarily set a cut-off value of 1 cm to define a substantial spleen enlargement. Even if a receiver operator characteristic curve could be built to derive the best cut-off value for the prediction of clinical outcomes, we preferred to use this simple value that can be evaluated easily and may reduce the risk of error caused by intraobserver or interobserver variability. We observed a significantly higher rate of appearance and growth of esophageal varices in patients showing a spleen enlargement compared with patients with unchanged spleen size. Moreover, among patients with compensated disease, patients showing an increased spleen size had a higher rate of first clinical decompensation as compared with patients with unchanged spleen size. When we tested spleen enlargement in a multivariate analysis including the most frequently recognized prognostic indicators, 15 it showed an independent predictive value both for variceal formation and growth, and for first clinical decompensation in compensated patients. In this setting, spleen enlargement determined a 4-fold increase in the risk of first clinical decompensation.

1134 BERZIGOTTI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6, No. 10 Portal hypertension is the major determinant of the formation of esophageal varices 20 and of decompensation in compen - varices detection? Hepatology 2001;33:333 338. cirrhosis should undergo endoscopic screening for esophageal sated patients, as recently shown in a large series of patients in8. Chalasani N, Imperiale TF, Ismail A, et al. Predictors of large whom hepatic venous pressure gradient (HVPG) was measured. 21,22 1999;94:3285 3291. esophageal varices in patients with cirrhosis. Am J Gastroenterol Based on the results of our study, we hypothesize 9. Thomopoulos KC, Labropoulou-Karatza C, Mimidis KP, et al. Noninvasive predictors of the presence of large oesophageal varices that spleen enlargement can be considered an equivalent of an increase in portal pressure, leading to an increased risk of in patients with cirrhosis. Dig Liver Dis 2003;35:473 478. clinical complications of portal hypertension. In a recent article, platelet count was studied in a cohort of patients with compensated cirrhosis and portal hypertension undergoing repeated measurements of HVPG and endoscopic screening of esophageal varices. 23 The investigators found that patients who developed esophageal varices had a significantly bleeding and nonbleeding esophageal varices. Am J Gastroenterol 1988;83:58 63. greater decrease in platelet count over time than patients who did not develop varices, suggesting that the dynamic change in 12. Ito K, Mitchell DG, Hann HW, et al. Compensated cirrhosis due to viral hepatitis: using MR imaging to predict clinical progression. platelet count can be a more reliable predictor of the occurrence AJR Am J Roentgenol 1997;169:801 805. of varices than platelet count at baseline. When we analyzed platelet count in our series, we observed that it decreased significantly during the follow-up period only in patients who showed an increase of spleen size at 1 year and who were at 14. Kamath PS, Wiesner RH, Malinchoc M, et al. A model to predict higher risk of developing complications of portal hypertension. survival in patients with end-stage liver disease. Hepatology In 2 previous studies 23,24 platelet count was shown to correlate 2001;33:464 470. with HVPG, but spleen size was not measured; conversely, 15. D Amico G, Garcia-Tsao G, Pagliaro L. Natural history and prognostic indicators of survival in cirrhosis: a systematic review of HVPG was not available in our patients, and data on direct correlations between portal pressure, spleen size, and platelet 118 studies. J Hepatol 2006;44:217 231. count are lacking. Our data show that in cirrhotic patients the platelet count correlates with spleen size, and that both spleen size and platelet count have a higher predictive value when analyzed as dynamic parameters. In the setting of patients with 17. Larson SM, Tuell SH, Moores KD, et al. Dimensions of the normal small varices at enrollment, both the increase in spleen size and adult spleen scan and prediction of spleen weight. J Nucl Med the decrease in platelet count at 1 year maintained an independent predictive value for variceal growth, suggesting that they18. Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal 1971;12:123 126. may reflect different aspects of the aggravation of portal hypertensive syndrome. 1981;27:213 218. hemorrhage by esophageal endoscopy. Gastrointest Endosc In conclusion, spleen enlargement on US follow-up evaluation is a simple, objective, and noninvasive parameter predict- 19. O Donohue J, Ng C, Catnach S, et al. Diagnostic value of Doppler ing the complications of portal hypertension. References 1. Ditchfield MR, Gibson RN, Donlan JD, et al. Duplex Doppler ultrasound signs of portal hypertension: relative diagnostic value of examination of paraumbilical vein, portal vein and spleen. Australas Radiol 1992;36:102 105. 2. Gibson PR, Gibson RN, Ditchfield MR, et al. Splenomegaly an insensitive sign of portal hypertension. Aust N Z J Med 1990;20: 771 774. 3. Bolognesi M, Merkel C, Sacerdoti D, et al. Role of spleen enlargement in cirrhosis with portal hypertension. Dig Liver Dis 2002; 34:144 150. 4. Shah SH, Hayes PC, Allan PL, et al. Measurement of spleen size and its relation to hypersplenism and portal hemodynamics in portal hypertension due to hepatic cirrhosis. Am J Gastroenterol 1996;91:2580 2583. 5. Berzigotti A, Schepis F, Garcia-Pagan JC. Monitoring treatment of cirrhosis and portal hypertension: noninvasive methods. In: Arroyo V, Sanchez-Fueyo A, Fernandez-Gomez J, et al, eds. Advances in therapy of liver diseases. Barcelona: Ars Medica, 2007:39 52. 6. Piscaglia F, Donati G, Cecilioni L, et al. Influence of the spleen on portal haemodynamics: a non-invasive study with Doppler ultrasound in chronic liver disease and haematological disorders. Scand J Gastroenterol 2002;37:1220 1227. 7. Schepis F, Camma C, Niceforo D, et al. Which patients with 10. Watanabe S, Hosomi N, Kitade Y, et al. Assessment of the presence and severity of esophagogastric varices by splenic index in patients with liver cirrhosis. J Comput Assist Tomogr 2000;24:788 794. 11. Medhat A, Iber FL, Dunne M, et al. Ultrasonographic findings with 13. Pugh RN, Murray-Lyon IM, Dawson JL, et al. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973; 60:646 649. 16. Giannini EG, Zaman A, Kreil A, et al. Platelet count/spleen diameter ratio for the noninvasive diagnosis of esophageal varices: results of a multicenter, prospective, validation study. Am J Gastroenterol 2006;101:2511 2519. assessment of the hepatic and portal vessels and ultrasound of the spleen in liver disease. Eur J Gastroenterol Hepatol 2004; 16:147 155. 20. Garcia-Tsao G, Groszmann RJ, Fisher RL, et al. Portal pressure, presence of gastroesophageal varices and variceal bleeding. Hepatology 1985;5:419 424. 21. Groszmann RJ, Garcia-Tsao G, Bosch J, et al. Beta-blockers to prevent gastroesophageal varices in patients with cirrhosis. N Engl J Med 2005;353:2254 2261. 22. Ripoll C, Groszmann R, Garcia-Tsao G, et al. Hepatic venous pressure gradient predicts clinical decompensation in patients with compensated cirrhosis. Gastroenterology 2007;133:481 488. 23. Qamar AA, Grace ND, Groszmann RJ, et al. Platelet count is not a predictor of the presence or development of gastroesophageal varices in cirrhosis. Hepatology 2008;47:153 159. 24. Dittrich S, de Mattos AA, Cheinquer H, et al. [Correlation between platelet blood levels and the hepatic venous pressure gradient among patients with cirrhosis]. Arq Gastroenterol 2005;42:35 40. Address requests for reprints to: Annalisa Berzigotti, MD, Dipartimento di Medicina Interna, Cardioangiologia, Epatologia, Università di Bologna, Policlinico S. Orsola-Malpighi, Via Albertoni 15, 40138 Bologna, Italy. e-mail: aberzigo@clinic.ub.es; fax: (39) 051-6362210. Supported in part by a grant from the Dipartimento di Medicina Interna, Cardioangiologia, Epatologia, Università di Bologna, Bologna, Italy. Dr A. Berzigotti receives financial support from a junior investigator research grant from the Dipartimento di Medicina Interna, Cardioangiologia, Epatologia, Università di Bologna, Italy.