bs_bs_banner
doi:10.1111/jgh.12658
H E PAT O L O G Y
Major predictors of portal hypertensive enteropathy in patients with liver cirrhosis Taiki Aoyama,* Shiro Oka,† Hiroshi Aikata,* Atsushi Igawa,* Makoto Nakano,* Noriaki Naeshiro,* Shigeto Yoshida,† Shinji Tanaka† and Kazuaki Chayama* *Department of Gastroenterology and Metabolism, Graduate School of Biomedical Sciences, Hiroshima University, and †Department of Endoscopy, Hiroshima University Hospital, Hiroshima, Japan
Key words capsule endoscopy, liver cirrhosis, portal hypertensive enteropathy, portosystemic shunts. Accepted for publication 19 May 2014. Correspondence Dr Shiro Oka, Department of Endoscopy, Hiroshima University Hospital, 1-2-3 Kasumi, Minami-ku, Hiroshima 734-8551, Japan. Email: [email protected] Financial disclosure: The work described herein was supported by departmental resources only. The authors have no conflicts of interest to disclose.
Abstract Background and Aim: Portal hypertensive enteropathy (PHE) is acknowledged as a source of bleeding, and predicting its presence has become more important. We assessed PHE using capsule endoscopy (CE) and investigated factors that may predict its presence, including portosystemic shunts (PSs). Methods: We analyzed data from 134 consecutive patients with liver cirrhosis, from February 2009 to September 2013. All patients had undergone dynamic computed tomography and esophagogastroduodenoscopy before CE examination. The frequencies and types of PHE lesions, and the relationships between the presence of PHE and patients’ clinical characteristics were evaluated. The distribution of the lesions was also determined. Results: PHE was found in 91 (68%), erythema in 70 (52%), erosions in 25 (19%), angioectasia in 24 (18%), villous edema in 18 (13%), and varices in 10 (7%) patients. Most lesions were located in the jejunum. The clinical characteristics associated with the presence of PHE were a Child-Pugh grade of B or C (P = 0.0058), and the presence of PSs (P < 0.0001), ascites (P = 0.0017), portal thrombosis (P = 0.016), esophageal varices (P = 0.0017), and portal hypertensive gastropathy (P = 0.0029). The presence of PSs was an independent predictor of PHE (odds ratio [OR]: 3.15; 95% confidence interval [CI]: 1.27–7.95). Among the shunt types, left gastric vein (OR: 5.31; 95% CI: 1.97–17.0) and splenorenal shunts (OR: 4.26; 95% CI: 1.29–19.4) were independent predictors of PHE. Conclusion: PSs, especially left gastric vein and splenorenal shunts, appear to reliably predict the presence of PHE.
Introduction Portal hypertension (PHT) is a frequent complication of liver cirrhosis (LC) and is generally defined as a portal pressure gradient that is greater than 10 mmHg.1 PHT can cause a variety of pathologic changes along the entire gastrointestinal tract, from the esophagus to the anus, which manifest as varices, gastropathy, and enteropathy.2,3 Advances in small-bowel endoscopy are being accompanied by greater debate about the pathogenesis and clinical characteristics of portal hypertensive enteropathy (PHE) alongside varices and gastropathy. From a clinical perspective, the significance of PHE lesions relates to them being possible sources of bleeding.4–9 Capsule endoscopy (CE) is accepted worldwide as a noninvasive technique for examining the small bowel.10,11 Further, CE and double-balloon endoscopy are comparable in terms of their abilities to detect lesions in the small bowel.12–14 De Palma et al. conducted a CE-based study and found PHE in 67% of patients with LC.2 They defined PHE as the presence of inflammatory-like 124
abnormalities, which present as edematous, erythematous, granular, and friable lesions and/or vascular lesions, which present as cherry-red spots, telangiectasias, or angiodysplasia-like lesions and varices.2 Since PHE lesions cause bleeding and subsequent anemia, predicting the presence of PHE is important. Although some publications have reported on the prevalence and the predictors of PHE in patients with LC,2,15–17 the prevalence of PHE and its associated clinical characteristics remain controversial. Hepatocellular carcinoma (HCC) is one of the main complications in patients with LC, and computed tomography (CT) is one of the preferred surveillance tools used to detect the development of HCC in patients with LC. With regard to the surveillance interval, semi-annual and annual surveillance programs have been shown to improve the survival of patients with LC and HCC equally.18,19 CT scanning provides a wealth of information in addition to that relating to HCC. For example, CT scans powerfully evaluate the overall status of the portosystemic shunts (PSs).20 Detecting these shunts is important to precisely evaluate hepatic hemodynamics and disease severity, and to extend the range of PHT indicators.
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Aoyama et al.
Portal pressure is important in defining the prognoses of patients with LC. However, measuring portal pressure is invasive and predicting portal pressure noninvasively is challenging. These factors emphasize the need for simple, clinically useful predictors of PHT and PHE. Therefore, we retrospectively conducted a CE-based study of patients with LC to assess PHE and explore factors that may predict PHE, including PSs.
Methods Patient population. One hundred and thirty-four consecutive patients with LC (78 males and 56 females), with a mean age of 66.7 years (range: 27–88 years) who underwent CE at Hiroshima University Hospital between February 2009 and September 2013, comprised the study population. CE was indicated if bleeding from the small bowel and/or iron deficiency anemia with hemoglobin levels of ≤12.0 g/dL was suspected. LC was diagnosed on the basis of a patient’s medical history and indicative physical findings, including vascular spiders, palmar erythema, or leg edema; laboratory findings, including impaired synthetic liver function and thrombocytopenia; upper abdominal radiographic characteristics, for example, a nodular appearance, an irregular liver contour, splenomegaly, and/or the presence of ascites; and/or liver biopsy assessments in which LC was determined as F4 fibrosis using the new Inuyama classification system to score fibrosis. Liver disease was of a viral etiology in 102 patients (hepatitis C virus: 76 patients, hepatitis B virus: 26 patients) and of a non-viral etiology in 32 patients (excessive alcohol consumption: 13 patients, non-alcoholic steatohepatitis: 7 patients, autoimmune hepatitis: 3 patients, and other causes: 9 patients). The Child-Pugh system was used to grade the severity of liver disease as follows: grade A = 61 patients, grade B = 62 patients, and grade C = 11 patients. The clinical characteristics of the patients are presented in Table 1. Patients who used either non-steroidal anti-inflammatory drugs or β-blockers were excluded from the study. All patients had undergone dynamic CT scanning and esophagogastroduodenoscopy (EGD) before they underwent CE examination. This study conformed with the guidelines of the Declaration of Helsinki (Sixth revision, 2008) as reflected by the a priori approval of the study by the Institutional Review Board at Hiroshima University Hospital which also granted permission for us to access the patients’ information.
CE. CE was performed using a PillCam SB/SB2 video capsule (Given Imaging Ltd, Yokneam, Israel). The sensor arrays were attached to the patients’ abdomens, the data recorders were attached to belts around the patients’ waists, and then the capsules were swallowed with a solution of dimethicone after a 12-h overnight fast. Once patients had swallowed the capsules, they were able to participate in normal activities. After 8 h, the sensor arrays and recording devices were removed, and the images were analyzed using Rapid Reader 6 software on a RAPID 5 or 6.5 workstation (Given Imaging Ltd). The images captured by CE were reviewed and interpreted by two endoscopists who had reviewed images from more than 200 patients. Diagnoses were reached by agreement.
Prediction of enteropathy
Table 1 Baseline demographic and clinical characteristics of patients with and without portal hypertensive enteropathy PHE
Demographic characteristic Age ≥ 65 years Male gender Clinical characteristic Child-Pugh grade B or C Viral etiology of cirrhosis§ CT findings PSs Ascites HCC Splenomegaly¶ Portal thrombosis†† Endoscopic findings‡‡ EVs GVs PHG
P-value
Present† n = 91
Absent† n = 43
54 (59) 50 (50)
24 (56) 28 (65)
0.70 0.27
57 (63) 66 (73)
16 (37) 36 (84)
0.0058‡ 0.16
73 (80) 45 (49) 57 (63) 55 (60) 19 (21)
19 (44) 9 (21) 24 (56) 22 (51) 2 (5)
< 0.0001 0.0017 0.45 0.31 0.016
62 (68) 26 (29) 36 (40)
17 (40) 9 (21) 6 (14)
0.0017 0.35 0.0029
†
Data are expressed as numbers, n, (%) of patients. Versus a Child-Pugh grade of A. § A viral etiology includes infection with hepatitis B virus and hepatitis C virus. A non-viral etiology includes alcoholic hepatitis, non-alcoholic steatohepatitis, autoimmune hepatitis, and others. ¶ Splenomegaly was considered when the splenic area was > 40 cm2. Splenomegaly absence includes patients who had undergone splenectomy. †† Portal thrombosis included portal vein tumor thrombosis. ‡‡ Determined by esophagogastroduodenoscopy. CT, computed tomography; EVs, esophageal varices; GVs, gastric varices; HCC, hepatocellular carcinoma; PHE, portal hypertensive enteropathy; PHG, portal hypertensive gastropathy; PSs, portosystemic shunts. ‡
Study measures. The frequencies and types of PHE lesions were investigated using CE, and Figure 1 presents CE images of PHE lesions. Moreover, the locations of the PHE lesions were determined in each patient. The entire small bowel was divided into three parts, based on the CE transit times and in accordance with the report by Li et al.21 The first two parts were considered as the jejunum, and the last part was considered as the ileum. The relationships between PHE and the patients’ clinical characteristics, which included age, sex, liver function, etiology of cirrhosis, PSs, ascites, HCC, splenomegaly, portal thrombosis, esophageal varices (EVs), gastric varices (GVs), and portal hypertensive gastropathy (PHG), were evaluated. The findings from EGD performed within 1 month and dynamic CT performed within 3 months of the CE examinations were used as reference points. CT examinations were performed using the high-quality scanning mode and a 1.25-mm slice thickness. PSs were evaluated using dynamic CT during the portal venous phase and coronal CT imagery. Left gastric veins, short gastric veins, posterior gastric veins, paraesophageal veins, splenorenal shunts, and paraumbilical veins with diameters greater than 3 mm were defined as PSs (Figure. 2).
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
125
Prediction of enteropathy
T Aoyama et al.
a
b
d
e
c
Figure 1 Capsule endoscopic images of portal hypertensive enteropathy in patients with liver cirrhosis. (a) Erythema, (b) erosions, (c) angioectasia, (d) varices, and (e) villous edema.
Statistical analysis. Comparisons were performed using Student’s t-test for quantitative data and the chi-square test for categorical data. Yates’ correction or Fisher’s exact test were used when required. All tests were two-sided, and a P-value < 0.05 was considered statistically significant. Predictive factors that were determined to be significant by univariate analysis were then subjected to multivariate analyses. The impacts of the clinical variables on PHE were estimated by calculating the odds ratios (OR) and the 95% confidence intervals (95% CI) using logistic regression analyses. The statistical software package JMP10 (SAS Institute Inc., Cary, NC, USA) was used for all calculations.
Results All capsules were evacuated, and this was confirmed for all patients. PHE was found in 91 (68%) patients, and 70 (52%) patients had erythema, 25 (19%) patients had erosions, 24 (18%) patients had angioectasia, 18 (13%) patients had villous edema, and 10 (7%) patients had varices. Erythema was present in 244 lesions (214 lesions in the jejunum and 30 lesions in the ileum). Erosions were present in 88 lesions (78 lesions in the jejunum and 10 lesions in the ileum). Angioectasia was noted in 45 lesions (43 lesions in the jejunum and 2 lesions in the ileum). Varices were observed in 15 lesions (10 lesions in the jejunum and 5 lesions in 126
the ileum). Villous edema was observed in the whole small bowel in 5 patients, the jejunum alone in 12 patients, and the ileum alone in 1 patient. The clinical characteristics of the cirrhotic patients with and without PHE are compared in Table 1. The clinical characteristics significantly associated with PHE were a Child-Pugh grade of B or C (vs grade A, P = 0.0058), and the presence of PSs (vs no PSs, P < 0.0001), ascites (vs no ascites, P = 0.0017), portal thrombosis (vs no portal thrombosis, P = 0.016), EVs (vs no EVs, P = 0.0017), and PHG (vs no PHG, P = 0.0029). The Child-Pugh classification is complex and includes an ascites index. Therefore, in the multivariate analysis, we incorporated PSs, ascites, portal thrombosis, EVs, and PHG, which were related to the prevalence of PHE according to the univariate analysis. Subsequent multivariate analysis (Table 2) determined that the presence of PSs was an independent predictor of PHE (OR: 3.15; 95% CI: 1.27–7.95). Table 3 shows the details of the types of shunt found in the cirrhotic patients with and without PHE and their subsequent multivariate analysis. The shunt types significantly associated with PHE on univariate analysis were left gastric vein (vs no left gastric vein, P = 0.00068), paraesophageal vein (vs no paraesophageal vein, P = 0.029), and splenorenal shunts (vs no splenorenal shunt, P = 0.03). Subsequent multivariate analysis determined that left gastric vein (OR: 5.31; 95% CI: 1.97–17.0) and splenorenal shunts
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Aoyama et al.
Prediction of enteropathy
a
b
c
d
e
f
Figure 2 Portosystemic shunts depicted by dynamic computed tomography (CT) during the portal venous phase, using coronal imagery. The arrows on the images depict the (a) left gastric vein, (b) short gastric vein, (c) posterior gastric vein, (d) paraesophageal vein, (e) splenorenal shunt, and (f) paraumbilical vein.
Table 2
Multivariate analysis of the clinical characteristics associated with portal hypertensive enteropathy
Clinical characteristic
Reference
Odds ratio
95% CI
P-value
PSs Ascites Portal thrombosis† EVs PHG
PSs absent Ascites absent Portal thrombosis absent EVs absent PHG absent
3.15 1.82 3.50 1.25 2.03
1.27–7.95 0.67–5.03 0.87–23.6 0.48–3.15 0.70–6.47
0.013 0.23 0.081 0.64 0.20
†
Portal thrombosis includes portal vein tumor thrombosis. CI, confidence interval; PSs, portosystemic shunts; EVs, esophageal varices; PHG, portal hypertensive gastropathy.
(OR: 4.26; 95% CI: 1.29–19.4) were independent predictors of PHE. Each type of PHE lesion was then evaluated based on the presence or absence of PSs (Table 4). Inflammatory-like abnormalities, in particular erythema and villous edema, were associated with the presence of PSs. Vascular lesions were not associated with the presence of PSs. Incidentally, with regard to the relationship between inflammatory-like abnormalities and hepatitis markers, we found that the mean (± standard deviation [SD]) levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) did not differ statistically between patients with inflammatory-like abnormalities and those without inflammatory-like abnormalities (47.8 [ ± 34.5] IU/L vs 46.5 [ ± 23.2] IU/L, respectively, P = 0.81), (34.8 [ ± 26.0] IU/L vs 38.3 [ ± 28.9] IU/L, respectively, P = 0.47). Patients with hepatitis C had inflammatory-like abnormalities and vascular lesions at a prevalence of 63% (48/76) and 20% (15/76), respectively. Patients with hepatitis B had inflammatory-like abnormalities and vascular lesions at a prevalence of 58% (15/26) and 31% (8/26), respectively. There was no difference between patients
with hepatitis C and those with hepatitis B in relation to PHE manifestation (P = 0.46). Levels of hemoglobin, serum iron, and serum ferritin were evaluated from the laboratory analyses. The mean (± SD) levels of serum ferritin did not differ statistically between patients with PHE and those without PHE (57.3 [ ± 74.0] ng/ml vs 99.8 [ ± 193] ng/dL, respectively, P = 0.086), but a comparison of patients with PHE and those without PHE showed that the mean (± SD) levels of hemoglobin (9.7 [ ± 1.9] g/dL vs 10.5 [ ± 2.4] g/dL, respectively, P = 0.040) and the mean (± SD) levels of serum iron (57.3 [ ± 41.2] μg/mL vs 84.0 [ ± 80.1] μg/mL, respectively, P = 0.018) were significantly lower in patients with PHE.
Discussion With progress in small-bowel endoscopy, there has been a greater focus on the pathogenesis and clinical characteristics of PHE because of its potential to cause considerable bleeding. Both the
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
127
Prediction of enteropathy
T Aoyama et al.
Table 3 Shunt types in patients with cirrhosis categorized according to the presence or absence of portal hypertensive enteropathy, and their multivariate analysis Shunt type
Left gastric vein Short gastric vein Posterior gastric vein Paraesophageal vein Splenorenal shunts Paraumbilical vein
PHE
Univariate analysis
Multivariate analysis
Present† n = 91
Absent† n = 43
P-value
Odds ratio
95% CI
P-value
39 (43) 1 (1) 17 (19) 28 (31) 21 (23) 8 (9)
5 (17) 0 (0) 8 (19) 5 (17) 3 (7) 5 (12)
0.00068 1 1 0.029 0.03 0.84
5.31 — — 2.06 4.26 —
1.97–17.0 — — 0.71–6.86 1.29–19.4 —
0.0006 — — 0.19 0.02 —
†
Data are expressed as numbers, n, (%) of patients. —, data not shown; CI, confidence interval; PHE, portal hypertensive enteropathy.
Table 4 Portal hypertensive enteropathy lesions in patients with and without portosystemic shunts PHE lesion
Inflammatory-like abnormalities Erythema Erosions Villous edema Vascular lesions Angioectasia Varices
PSs
P-value
Present† n = 92
Absent† n = 42
68 (74) 56 (61) 20 (22) 16 (17) 26 (28) 20 (22) 7 (8)
16 (38) 14 (33) 5 (12) 2 (5) 7 (17) 4 (10) 3 (7)
< 0.0001 0.0031 0.18 0.047 0.15 0.087 0.92
†
Data are expressed as numbers, n, (%) of patients. PHE, portal hypertensive enteropathy; PSs, portosystemic shunts.
clinical characteristics of PHE and the clinical manifestations that have associations with PHE have been reported. To date, these include a Child-Pugh grade C, large EVs, PHG, portal hypertensive colopathy, and ascites.2,9,15,16 In addition to these factors, a complex scoring system based on CT findings has been reported as a useful predictor of PHE.17 However, we have been seeking more definitive factors that can be used as predictors of PHE. In this study, we found clinical characteristics associated with PHE that are similar to those previously reported, including a Child-Pugh grade of B or C, PSs, ascites, portal thrombosis, EVs, and PHG. Interestingly, however, only PSs were found to independently predict PHE upon multivariate analysis. This simple and appealing factor is easy to observe during an annual CT scan. This led us to investigate the significance of PSs and their clinical associations with PHE. PSs develop in an environment where the portal pressure is higher than 10 mmHg, and they act as bypasses to compensate for PHT.1 The presence of PSs indicates the severity of PHT and the exacerbation of EVs.22,23 In fact, PSs reflected the status of liver function in our study because more patients with PSs had a ChildPugh grade of B or C (n = 62; 46%) than a Child-Pugh grade of A (n = 30; 22%) (P = 0.00001). Although this finding is useful, it is important to distinguish among the different types of PSs. Zardi et al. reported that each shunt differs in its ability to alleviate portal pressure, and that the absence of PSs or the presence of paraum128
bilical veins could help to identify a more favorable clinical course in patients with LC.22 Conversely, they also reported that left gastric vein and splenorenal shunts frequently indicate more severe disease.22 In our study, left gastric vein and splenorenal shunts were independent predictors of PHE. The left gastric vein plays a key role as a feeder of EVs. Indeed, the driving force for the development of EVs is a high blood flow velocity in the left gastric vein, reflecting an increase in portal pressure.24,25 Kumamoto et al.26 proposed that since the prevalence of splenorenal shunts is similar to that of encephalopathy, the presence of a splenorenal shunt may be a prognostic factor for a variety of diseases in addition to encephalopathy because it exacerbates the decline in liver function by stealing the portal flow. During longterm monitoring, they demonstrated worsening survival rates in patients with splenorenal shunts.26 PHE frequently occurs in patients with cirrhosis and is as common as PHG. Barakat et al. reported that endoscopic duodenopathy was significantly higher in patients with chronic liver disease and PHT accompanied by severe gastropathy and that it caused gastrointestinal bleeding in 9.5% of patients.27 Chronic anemia is frequently observed in patients with LC, and although the mechanism underlying the development of anemia in patients with LC is multifactorial, acute or chronic blood loss may contribute to its development. In this study, hemoglobin and serum iron levels were significantly lower in patients with PHE than in those without PHE. Whether the presence of PHE itself was the reason for the low level of hemoglobin remains to be investigated in a future study. There would certainly have been patients in which PHE was the reason for their anemia. However, other patients might have been experiencing chronic blood loss from the esophagus/stomach or small bowel, or a combination of both. EVs and PHG are reported to hemorrhage at an estimated yearly rate of 5–15% and 8–25%, respectively.28,29 No reports have been published, as far as we are aware, that have stated the estimated yearly rate of hemorrhage in patients with PHE. Currently, small-bowel angioectasias and varices are thought to account for the unexplained blood losses in patients with PHT,4,6 and some of these could be suitable candidates for endoscopic treatment.15 Furthermore, we should remember that even villous edema appears to be susceptible to bleeding.7 When the PHE lesions were categorized into two groups in this study in accordance with De Palma et al.,2 inflammatory-like abnormalities predominated among our patients (n = 84, 63%), Journal of Gastroenterology and Hepatology 30 (2015) 124–130
© 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Aoyama et al.
followed by vascular lesions (n = 33, 25%). When these lesions were evaluated based on the presence of PSs, the PHE lesions that were associated with PSs were inflammatory-like abnormalities, including erythema and villous edema, rather than vascular lesions. We did not assess the hepatic inflammatory markers, AST and ALT, as predictors of these lesions. Congestion in the portal system is considered the main mucosal alteration in the small bowel that occurs in a similar manner as PHG.30,31 Although the development of vascular lesions is considered similar to that of PSs, we could not demonstrate a close relationship between PSs and vascular lesions, perhaps because of the multifactorial processes that underlie PHE development that may manifest themselves diversely. Similarly, GVs were not significantly related to the presence of PHE, probably because of the anatomical and hemodynamic diversity presented by GVs.32 In relation to whether inflammatory-like abnormalities are more dependent on portal pressure, Takahashi et al. showed that edema of the small bowel was strongly correlated with portal venous pressure.33 Furthermore, Matsushita et al. reported that reddened lesions and edema in the small bowel were attenuated once the portal pressure was reduced by a transjugular intrahepatic portosystemic shunt, but that the number of angioectasias did not change because of the reduction in portal pressure.34 Further debate is required regarding the treatment of PHE lesions. Our study has some limitations. First, we did not compare our findings with those from a control healthy group because of the retrospective nature of the study. Second, the study included patients from a single center only. Therefore, a large-scale study is needed to address these limitations. In relation to the consequences of PHT, PSs appear to reliably predict PHE and we can conclude that PHE presence can be predicted. Moreover, CE should be considered for patients with LC accompanied by PSs, especially where there are left gastric vein and splenorenal shunts.
References 1 Martínez-Noguera A, Montserrat E, Torrubia S, Villalba J. Doppler in hepatic cirrhosis and chronic hepatitis. Semin. Ultrasound CT MR 2002; 23: 19–36. 2 De Palma GD, Rega M, Masone S et al. Mucosal abnormalities of the small bowel in patients with cirrhosis and portal hypertension: a capsule endoscopy study. Gastrointest. Endosc. 2005; 62: 529–34. 3 Viggiano TR, Gostout CJ. Portal hypertensive intestinal vasculopathy: a review of the clinical, endoscopic, and histopathologic features. Am. J. Gastroenterol. 1992; 87: 944–54. 4 Akyuz F, Pinarbasi B, Ermis F et al. Is portal hypertensive enteropathy an important additional cause of blood loss in portal hypertensive patients? Scand. J. Gastroenterol. 2010; 45: 1497–502. 5 Yamada A, Watabe H, Obi S et al. Surveillance of small intestinal abnormalities in patients with hepatocellular carcinoma: a prospective capsule endoscopy study. Dig. Endosc. 2011; 23: 124–9. 6 Tang SJ, Zanati S, Dubcenco E et al. Diagnosis of small-bowel varices by capsule endoscopy. Gastrointest. Endosc. 2004; 60: 129–35. 7 Higaki N, Matsui H, Imaoka H et al. Characteristic endoscopic features of portal hypertensive enteropathy. J. Gastroenterol. 2008; 43: 327–31.
Prediction of enteropathy
8 Raju GS, Gerson L, Das A, Lewis B. American Gastroenterological Association (AGA) Institute technical review on obscure gastrointestinal bleeding. Gastroenterology 2007; 133: 1697–717. 9 Kodama M, Uto H, Numata M et al. Endoscopic characterization of the small bowel in patients with portal hypertension evaluated by double balloon endoscopy. J. Gastroenterol. 2008; 43: 589–96. 10 Roberts-Thomson IC, Singh R, Teo E, Nguyen NQ, Lidums I. The future of endoscopy. J. Gastroenterol. Hepatol. 2010; 25: 1051–7. 11 Imagawa H, Oka S, Tanaka S et al. Improved visibility of lesions of the small intestine via capsule endoscopy with computed virtual chromoendoscopy. Gastrointest. Endosc. 2011; 73: 299–306. 12 Fukumoto A, Tanaka S, Shishido T, Takemura Y, Oka S, Chayama K. Comparison of detectability of small-bowel lesions between capsule endoscopy and double-balloon endoscopy for patients with suspected small-bowel disease. Gastrointest. Endosc. 2009; 69: 857–65. 13 Shishido T, Oka S, Tanaka S et al. Diagnostic yield of capsule endoscopy vs. double-balloon endoscopy for patients who have undergone total enteroscopy with obscure gastrointestinal bleeding. Hepatogastroenterology 2012; 59: 955–9. 14 Teshima CW, Kuipers EJ, van Zanten SV, Mensink PB. Double balloon enteroscopy and capsule endoscopy for obscure gastrointestinal bleeding: an updated meta-analysis. J. Gastroenterol. Hepatol. 2011; 26: 796–801. 15 Aoyama T, Oka S, Aikata H et al. Small bowel abnormalities in patients with compensated liver cirrhosis. Dig. Dis. Sci. 2013; 58: 1390–6. 16 Goulas S, Triantafyllidou K, Karagiannis S et al. Capsule endoscopy in the investigation of patients with portal hypertension and anemia. Can. J. Gastroenterol. 2008; 22: 469–74. 17 Jeon SR, Kim JO, Kim JB et al. Portal hypertensive enteropathy diagnosed by capsule endoscopy in cirrhotic patients: a nationwide multicenter study. Dig. Dis. Sci. 2014; 59: 1036–41. 18 Trevisani F, De Notariis S, Rapaccini G et al. Semiannual and annual surveillance of cirrhotic patients for hepatocellular carcinoma: effects on cancer stage and patient survival (Italian experience). Am. J. Gastroenterol. 2002; 97: 734–44. 19 El-Serag HB, Marrero JA, Rudolph L, Reddy KR. Diagnosis and treatment of hepatocellular carcinoma. Gastroenterology 2008; 134: 1752–63. 20 Weishaupt D, Pfammatter T, Hilfiker PR, Wolfensberger U, Marincek B. Detecting bleeding duodenal varices with multislice helical CT. AJR Am. J. Roentgenol. 2002; 178: 399–401. 21 Li X, Chen H, Dai J, Gao Y, Ge Z. Predictive role of capsule endoscopy on the insertion route of double-balloon enteroscopy. Endoscopy 2009; 41: 762–6. 22 Zardi EM, Uwechie V, Caccavo D et al. Portosystemic shunts in a large cohort of patients with liver cirrhosis: detection rate and clinical relevance. J. Gastroenterol. 2009; 44: 76–83. 23 Nagaoki Y, Aikata H, Kobayashi T et al. Risk factors for the exacerbation of esophageal varices or portosystemic encephalopathy after sustained virological response with IFN therapy for HCV-related compensated cirrhosis. J. Gastroenterol. 2013; 48: 847–55. 24 Hino S, Kakutani H, Ikeda K et al. Hemodynamic assessment of the left gastric vein in patients with esophageal varices with color Doppler EUS: factors affecting development of esophageal varices. Gastrointest. Endosc. 2002; 55: 512–7. 25 Bosch J, Abraldes JG, Berzigotti A, Garcia-Pagan JC. Portal hypertension and gastrointestinal bleeding. Semin. Liver Dis. 2008; 28: 3–25. 26 Kumamoto M, Toyonaga A, Inoue H et al. Long-term results of balloon-occluded retrograde transvenous obliteration for gastric fundal varices: hepatic deterioration links to portosystemic shunt syndrome. J. Gastroenterol. Hepatol. 2010; 25: 1129–35.
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
129
Prediction of enteropathy
T Aoyama et al.
27 Barakat M, Mostafa M, Mahran Z, Soliman AG. Portal hypertensive duodenopathy: clinical, endoscopic, and histopathologic profiles. Am. J. Gastroenterol. 2007; 102: 2793–802. 28 North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices. Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices. A prospective multicenter study. N. Engl. J. Med. 1988; 319: 983–9. 29 D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. Hepatology 1995; 22: 332–54. 30 Toyoda H, Nakao M, Ogura Y et al. Congestion of superior mesenteric veins and small bowel mucosal injury after endoscopic treatment of esophageal varices in patients with portal hypertension. Dig. Dis. Sci. 2001; 46: 2353–9.
130
31 Misra V, Misra SP, Dwivedi M, Gupta SC. Histomorphometric study of portal hypertensive enteropathy. Am. J. Clin. Pathol. 1997; 108: 652–7. 32 Saad WE. Vascular anatomy and the morphologic and hemodynamic classifications of gastric varices and spontaneous portosystemic shunts relevant to the BRTO procedure. Tech. Vasc. Interv. Radiol. 2013; 16: 60–100. 33 Takahashi Y, Fujimori S, Narahara Y et al. Small intestinal edema had the strongest correlation with portal venous pressure amongst capsule endoscopy findings. Digestion 2012; 86: 48–54. 34 Matsushita Y, Narahara Y, Fujimori S et al. Effects of transjugular intrahepatic portosystemic shunt on changes in the small bowel mucosa of cirrhotic patients with portal hypertension. J. Gastroenterol. 2013; 48: 633–9.
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
doi:10.1111/jgh.12658
H E PAT O L O G Y
Major predictors of portal hypertensive enteropathy in patients with liver cirrhosis Taiki Aoyama,* Shiro Oka,† Hiroshi Aikata,* Atsushi Igawa,* Makoto Nakano,* Noriaki Naeshiro,* Shigeto Yoshida,† Shinji Tanaka† and Kazuaki Chayama* *Department of Gastroenterology and Metabolism, Graduate School of Biomedical Sciences, Hiroshima University, and †Department of Endoscopy, Hiroshima University Hospital, Hiroshima, Japan
Key words capsule endoscopy, liver cirrhosis, portal hypertensive enteropathy, portosystemic shunts. Accepted for publication 19 May 2014. Correspondence Dr Shiro Oka, Department of Endoscopy, Hiroshima University Hospital, 1-2-3 Kasumi, Minami-ku, Hiroshima 734-8551, Japan. Email: [email protected] Financial disclosure: The work described herein was supported by departmental resources only. The authors have no conflicts of interest to disclose.
Abstract Background and Aim: Portal hypertensive enteropathy (PHE) is acknowledged as a source of bleeding, and predicting its presence has become more important. We assessed PHE using capsule endoscopy (CE) and investigated factors that may predict its presence, including portosystemic shunts (PSs). Methods: We analyzed data from 134 consecutive patients with liver cirrhosis, from February 2009 to September 2013. All patients had undergone dynamic computed tomography and esophagogastroduodenoscopy before CE examination. The frequencies and types of PHE lesions, and the relationships between the presence of PHE and patients’ clinical characteristics were evaluated. The distribution of the lesions was also determined. Results: PHE was found in 91 (68%), erythema in 70 (52%), erosions in 25 (19%), angioectasia in 24 (18%), villous edema in 18 (13%), and varices in 10 (7%) patients. Most lesions were located in the jejunum. The clinical characteristics associated with the presence of PHE were a Child-Pugh grade of B or C (P = 0.0058), and the presence of PSs (P < 0.0001), ascites (P = 0.0017), portal thrombosis (P = 0.016), esophageal varices (P = 0.0017), and portal hypertensive gastropathy (P = 0.0029). The presence of PSs was an independent predictor of PHE (odds ratio [OR]: 3.15; 95% confidence interval [CI]: 1.27–7.95). Among the shunt types, left gastric vein (OR: 5.31; 95% CI: 1.97–17.0) and splenorenal shunts (OR: 4.26; 95% CI: 1.29–19.4) were independent predictors of PHE. Conclusion: PSs, especially left gastric vein and splenorenal shunts, appear to reliably predict the presence of PHE.
Introduction Portal hypertension (PHT) is a frequent complication of liver cirrhosis (LC) and is generally defined as a portal pressure gradient that is greater than 10 mmHg.1 PHT can cause a variety of pathologic changes along the entire gastrointestinal tract, from the esophagus to the anus, which manifest as varices, gastropathy, and enteropathy.2,3 Advances in small-bowel endoscopy are being accompanied by greater debate about the pathogenesis and clinical characteristics of portal hypertensive enteropathy (PHE) alongside varices and gastropathy. From a clinical perspective, the significance of PHE lesions relates to them being possible sources of bleeding.4–9 Capsule endoscopy (CE) is accepted worldwide as a noninvasive technique for examining the small bowel.10,11 Further, CE and double-balloon endoscopy are comparable in terms of their abilities to detect lesions in the small bowel.12–14 De Palma et al. conducted a CE-based study and found PHE in 67% of patients with LC.2 They defined PHE as the presence of inflammatory-like 124
abnormalities, which present as edematous, erythematous, granular, and friable lesions and/or vascular lesions, which present as cherry-red spots, telangiectasias, or angiodysplasia-like lesions and varices.2 Since PHE lesions cause bleeding and subsequent anemia, predicting the presence of PHE is important. Although some publications have reported on the prevalence and the predictors of PHE in patients with LC,2,15–17 the prevalence of PHE and its associated clinical characteristics remain controversial. Hepatocellular carcinoma (HCC) is one of the main complications in patients with LC, and computed tomography (CT) is one of the preferred surveillance tools used to detect the development of HCC in patients with LC. With regard to the surveillance interval, semi-annual and annual surveillance programs have been shown to improve the survival of patients with LC and HCC equally.18,19 CT scanning provides a wealth of information in addition to that relating to HCC. For example, CT scans powerfully evaluate the overall status of the portosystemic shunts (PSs).20 Detecting these shunts is important to precisely evaluate hepatic hemodynamics and disease severity, and to extend the range of PHT indicators.
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Aoyama et al.
Portal pressure is important in defining the prognoses of patients with LC. However, measuring portal pressure is invasive and predicting portal pressure noninvasively is challenging. These factors emphasize the need for simple, clinically useful predictors of PHT and PHE. Therefore, we retrospectively conducted a CE-based study of patients with LC to assess PHE and explore factors that may predict PHE, including PSs.
Methods Patient population. One hundred and thirty-four consecutive patients with LC (78 males and 56 females), with a mean age of 66.7 years (range: 27–88 years) who underwent CE at Hiroshima University Hospital between February 2009 and September 2013, comprised the study population. CE was indicated if bleeding from the small bowel and/or iron deficiency anemia with hemoglobin levels of ≤12.0 g/dL was suspected. LC was diagnosed on the basis of a patient’s medical history and indicative physical findings, including vascular spiders, palmar erythema, or leg edema; laboratory findings, including impaired synthetic liver function and thrombocytopenia; upper abdominal radiographic characteristics, for example, a nodular appearance, an irregular liver contour, splenomegaly, and/or the presence of ascites; and/or liver biopsy assessments in which LC was determined as F4 fibrosis using the new Inuyama classification system to score fibrosis. Liver disease was of a viral etiology in 102 patients (hepatitis C virus: 76 patients, hepatitis B virus: 26 patients) and of a non-viral etiology in 32 patients (excessive alcohol consumption: 13 patients, non-alcoholic steatohepatitis: 7 patients, autoimmune hepatitis: 3 patients, and other causes: 9 patients). The Child-Pugh system was used to grade the severity of liver disease as follows: grade A = 61 patients, grade B = 62 patients, and grade C = 11 patients. The clinical characteristics of the patients are presented in Table 1. Patients who used either non-steroidal anti-inflammatory drugs or β-blockers were excluded from the study. All patients had undergone dynamic CT scanning and esophagogastroduodenoscopy (EGD) before they underwent CE examination. This study conformed with the guidelines of the Declaration of Helsinki (Sixth revision, 2008) as reflected by the a priori approval of the study by the Institutional Review Board at Hiroshima University Hospital which also granted permission for us to access the patients’ information.
CE. CE was performed using a PillCam SB/SB2 video capsule (Given Imaging Ltd, Yokneam, Israel). The sensor arrays were attached to the patients’ abdomens, the data recorders were attached to belts around the patients’ waists, and then the capsules were swallowed with a solution of dimethicone after a 12-h overnight fast. Once patients had swallowed the capsules, they were able to participate in normal activities. After 8 h, the sensor arrays and recording devices were removed, and the images were analyzed using Rapid Reader 6 software on a RAPID 5 or 6.5 workstation (Given Imaging Ltd). The images captured by CE were reviewed and interpreted by two endoscopists who had reviewed images from more than 200 patients. Diagnoses were reached by agreement.
Prediction of enteropathy
Table 1 Baseline demographic and clinical characteristics of patients with and without portal hypertensive enteropathy PHE
Demographic characteristic Age ≥ 65 years Male gender Clinical characteristic Child-Pugh grade B or C Viral etiology of cirrhosis§ CT findings PSs Ascites HCC Splenomegaly¶ Portal thrombosis†† Endoscopic findings‡‡ EVs GVs PHG
P-value
Present† n = 91
Absent† n = 43
54 (59) 50 (50)
24 (56) 28 (65)
0.70 0.27
57 (63) 66 (73)
16 (37) 36 (84)
0.0058‡ 0.16
73 (80) 45 (49) 57 (63) 55 (60) 19 (21)
19 (44) 9 (21) 24 (56) 22 (51) 2 (5)
< 0.0001 0.0017 0.45 0.31 0.016
62 (68) 26 (29) 36 (40)
17 (40) 9 (21) 6 (14)
0.0017 0.35 0.0029
†
Data are expressed as numbers, n, (%) of patients. Versus a Child-Pugh grade of A. § A viral etiology includes infection with hepatitis B virus and hepatitis C virus. A non-viral etiology includes alcoholic hepatitis, non-alcoholic steatohepatitis, autoimmune hepatitis, and others. ¶ Splenomegaly was considered when the splenic area was > 40 cm2. Splenomegaly absence includes patients who had undergone splenectomy. †† Portal thrombosis included portal vein tumor thrombosis. ‡‡ Determined by esophagogastroduodenoscopy. CT, computed tomography; EVs, esophageal varices; GVs, gastric varices; HCC, hepatocellular carcinoma; PHE, portal hypertensive enteropathy; PHG, portal hypertensive gastropathy; PSs, portosystemic shunts. ‡
Study measures. The frequencies and types of PHE lesions were investigated using CE, and Figure 1 presents CE images of PHE lesions. Moreover, the locations of the PHE lesions were determined in each patient. The entire small bowel was divided into three parts, based on the CE transit times and in accordance with the report by Li et al.21 The first two parts were considered as the jejunum, and the last part was considered as the ileum. The relationships between PHE and the patients’ clinical characteristics, which included age, sex, liver function, etiology of cirrhosis, PSs, ascites, HCC, splenomegaly, portal thrombosis, esophageal varices (EVs), gastric varices (GVs), and portal hypertensive gastropathy (PHG), were evaluated. The findings from EGD performed within 1 month and dynamic CT performed within 3 months of the CE examinations were used as reference points. CT examinations were performed using the high-quality scanning mode and a 1.25-mm slice thickness. PSs were evaluated using dynamic CT during the portal venous phase and coronal CT imagery. Left gastric veins, short gastric veins, posterior gastric veins, paraesophageal veins, splenorenal shunts, and paraumbilical veins with diameters greater than 3 mm were defined as PSs (Figure. 2).
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
125
Prediction of enteropathy
T Aoyama et al.
a
b
d
e
c
Figure 1 Capsule endoscopic images of portal hypertensive enteropathy in patients with liver cirrhosis. (a) Erythema, (b) erosions, (c) angioectasia, (d) varices, and (e) villous edema.
Statistical analysis. Comparisons were performed using Student’s t-test for quantitative data and the chi-square test for categorical data. Yates’ correction or Fisher’s exact test were used when required. All tests were two-sided, and a P-value < 0.05 was considered statistically significant. Predictive factors that were determined to be significant by univariate analysis were then subjected to multivariate analyses. The impacts of the clinical variables on PHE were estimated by calculating the odds ratios (OR) and the 95% confidence intervals (95% CI) using logistic regression analyses. The statistical software package JMP10 (SAS Institute Inc., Cary, NC, USA) was used for all calculations.
Results All capsules were evacuated, and this was confirmed for all patients. PHE was found in 91 (68%) patients, and 70 (52%) patients had erythema, 25 (19%) patients had erosions, 24 (18%) patients had angioectasia, 18 (13%) patients had villous edema, and 10 (7%) patients had varices. Erythema was present in 244 lesions (214 lesions in the jejunum and 30 lesions in the ileum). Erosions were present in 88 lesions (78 lesions in the jejunum and 10 lesions in the ileum). Angioectasia was noted in 45 lesions (43 lesions in the jejunum and 2 lesions in the ileum). Varices were observed in 15 lesions (10 lesions in the jejunum and 5 lesions in 126
the ileum). Villous edema was observed in the whole small bowel in 5 patients, the jejunum alone in 12 patients, and the ileum alone in 1 patient. The clinical characteristics of the cirrhotic patients with and without PHE are compared in Table 1. The clinical characteristics significantly associated with PHE were a Child-Pugh grade of B or C (vs grade A, P = 0.0058), and the presence of PSs (vs no PSs, P < 0.0001), ascites (vs no ascites, P = 0.0017), portal thrombosis (vs no portal thrombosis, P = 0.016), EVs (vs no EVs, P = 0.0017), and PHG (vs no PHG, P = 0.0029). The Child-Pugh classification is complex and includes an ascites index. Therefore, in the multivariate analysis, we incorporated PSs, ascites, portal thrombosis, EVs, and PHG, which were related to the prevalence of PHE according to the univariate analysis. Subsequent multivariate analysis (Table 2) determined that the presence of PSs was an independent predictor of PHE (OR: 3.15; 95% CI: 1.27–7.95). Table 3 shows the details of the types of shunt found in the cirrhotic patients with and without PHE and their subsequent multivariate analysis. The shunt types significantly associated with PHE on univariate analysis were left gastric vein (vs no left gastric vein, P = 0.00068), paraesophageal vein (vs no paraesophageal vein, P = 0.029), and splenorenal shunts (vs no splenorenal shunt, P = 0.03). Subsequent multivariate analysis determined that left gastric vein (OR: 5.31; 95% CI: 1.97–17.0) and splenorenal shunts
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Aoyama et al.
Prediction of enteropathy
a
b
c
d
e
f
Figure 2 Portosystemic shunts depicted by dynamic computed tomography (CT) during the portal venous phase, using coronal imagery. The arrows on the images depict the (a) left gastric vein, (b) short gastric vein, (c) posterior gastric vein, (d) paraesophageal vein, (e) splenorenal shunt, and (f) paraumbilical vein.
Table 2
Multivariate analysis of the clinical characteristics associated with portal hypertensive enteropathy
Clinical characteristic
Reference
Odds ratio
95% CI
P-value
PSs Ascites Portal thrombosis† EVs PHG
PSs absent Ascites absent Portal thrombosis absent EVs absent PHG absent
3.15 1.82 3.50 1.25 2.03
1.27–7.95 0.67–5.03 0.87–23.6 0.48–3.15 0.70–6.47
0.013 0.23 0.081 0.64 0.20
†
Portal thrombosis includes portal vein tumor thrombosis. CI, confidence interval; PSs, portosystemic shunts; EVs, esophageal varices; PHG, portal hypertensive gastropathy.
(OR: 4.26; 95% CI: 1.29–19.4) were independent predictors of PHE. Each type of PHE lesion was then evaluated based on the presence or absence of PSs (Table 4). Inflammatory-like abnormalities, in particular erythema and villous edema, were associated with the presence of PSs. Vascular lesions were not associated with the presence of PSs. Incidentally, with regard to the relationship between inflammatory-like abnormalities and hepatitis markers, we found that the mean (± standard deviation [SD]) levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) did not differ statistically between patients with inflammatory-like abnormalities and those without inflammatory-like abnormalities (47.8 [ ± 34.5] IU/L vs 46.5 [ ± 23.2] IU/L, respectively, P = 0.81), (34.8 [ ± 26.0] IU/L vs 38.3 [ ± 28.9] IU/L, respectively, P = 0.47). Patients with hepatitis C had inflammatory-like abnormalities and vascular lesions at a prevalence of 63% (48/76) and 20% (15/76), respectively. Patients with hepatitis B had inflammatory-like abnormalities and vascular lesions at a prevalence of 58% (15/26) and 31% (8/26), respectively. There was no difference between patients
with hepatitis C and those with hepatitis B in relation to PHE manifestation (P = 0.46). Levels of hemoglobin, serum iron, and serum ferritin were evaluated from the laboratory analyses. The mean (± SD) levels of serum ferritin did not differ statistically between patients with PHE and those without PHE (57.3 [ ± 74.0] ng/ml vs 99.8 [ ± 193] ng/dL, respectively, P = 0.086), but a comparison of patients with PHE and those without PHE showed that the mean (± SD) levels of hemoglobin (9.7 [ ± 1.9] g/dL vs 10.5 [ ± 2.4] g/dL, respectively, P = 0.040) and the mean (± SD) levels of serum iron (57.3 [ ± 41.2] μg/mL vs 84.0 [ ± 80.1] μg/mL, respectively, P = 0.018) were significantly lower in patients with PHE.
Discussion With progress in small-bowel endoscopy, there has been a greater focus on the pathogenesis and clinical characteristics of PHE because of its potential to cause considerable bleeding. Both the
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
127
Prediction of enteropathy
T Aoyama et al.
Table 3 Shunt types in patients with cirrhosis categorized according to the presence or absence of portal hypertensive enteropathy, and their multivariate analysis Shunt type
Left gastric vein Short gastric vein Posterior gastric vein Paraesophageal vein Splenorenal shunts Paraumbilical vein
PHE
Univariate analysis
Multivariate analysis
Present† n = 91
Absent† n = 43
P-value
Odds ratio
95% CI
P-value
39 (43) 1 (1) 17 (19) 28 (31) 21 (23) 8 (9)
5 (17) 0 (0) 8 (19) 5 (17) 3 (7) 5 (12)
0.00068 1 1 0.029 0.03 0.84
5.31 — — 2.06 4.26 —
1.97–17.0 — — 0.71–6.86 1.29–19.4 —
0.0006 — — 0.19 0.02 —
†
Data are expressed as numbers, n, (%) of patients. —, data not shown; CI, confidence interval; PHE, portal hypertensive enteropathy.
Table 4 Portal hypertensive enteropathy lesions in patients with and without portosystemic shunts PHE lesion
Inflammatory-like abnormalities Erythema Erosions Villous edema Vascular lesions Angioectasia Varices
PSs
P-value
Present† n = 92
Absent† n = 42
68 (74) 56 (61) 20 (22) 16 (17) 26 (28) 20 (22) 7 (8)
16 (38) 14 (33) 5 (12) 2 (5) 7 (17) 4 (10) 3 (7)
< 0.0001 0.0031 0.18 0.047 0.15 0.087 0.92
†
Data are expressed as numbers, n, (%) of patients. PHE, portal hypertensive enteropathy; PSs, portosystemic shunts.
clinical characteristics of PHE and the clinical manifestations that have associations with PHE have been reported. To date, these include a Child-Pugh grade C, large EVs, PHG, portal hypertensive colopathy, and ascites.2,9,15,16 In addition to these factors, a complex scoring system based on CT findings has been reported as a useful predictor of PHE.17 However, we have been seeking more definitive factors that can be used as predictors of PHE. In this study, we found clinical characteristics associated with PHE that are similar to those previously reported, including a Child-Pugh grade of B or C, PSs, ascites, portal thrombosis, EVs, and PHG. Interestingly, however, only PSs were found to independently predict PHE upon multivariate analysis. This simple and appealing factor is easy to observe during an annual CT scan. This led us to investigate the significance of PSs and their clinical associations with PHE. PSs develop in an environment where the portal pressure is higher than 10 mmHg, and they act as bypasses to compensate for PHT.1 The presence of PSs indicates the severity of PHT and the exacerbation of EVs.22,23 In fact, PSs reflected the status of liver function in our study because more patients with PSs had a ChildPugh grade of B or C (n = 62; 46%) than a Child-Pugh grade of A (n = 30; 22%) (P = 0.00001). Although this finding is useful, it is important to distinguish among the different types of PSs. Zardi et al. reported that each shunt differs in its ability to alleviate portal pressure, and that the absence of PSs or the presence of paraum128
bilical veins could help to identify a more favorable clinical course in patients with LC.22 Conversely, they also reported that left gastric vein and splenorenal shunts frequently indicate more severe disease.22 In our study, left gastric vein and splenorenal shunts were independent predictors of PHE. The left gastric vein plays a key role as a feeder of EVs. Indeed, the driving force for the development of EVs is a high blood flow velocity in the left gastric vein, reflecting an increase in portal pressure.24,25 Kumamoto et al.26 proposed that since the prevalence of splenorenal shunts is similar to that of encephalopathy, the presence of a splenorenal shunt may be a prognostic factor for a variety of diseases in addition to encephalopathy because it exacerbates the decline in liver function by stealing the portal flow. During longterm monitoring, they demonstrated worsening survival rates in patients with splenorenal shunts.26 PHE frequently occurs in patients with cirrhosis and is as common as PHG. Barakat et al. reported that endoscopic duodenopathy was significantly higher in patients with chronic liver disease and PHT accompanied by severe gastropathy and that it caused gastrointestinal bleeding in 9.5% of patients.27 Chronic anemia is frequently observed in patients with LC, and although the mechanism underlying the development of anemia in patients with LC is multifactorial, acute or chronic blood loss may contribute to its development. In this study, hemoglobin and serum iron levels were significantly lower in patients with PHE than in those without PHE. Whether the presence of PHE itself was the reason for the low level of hemoglobin remains to be investigated in a future study. There would certainly have been patients in which PHE was the reason for their anemia. However, other patients might have been experiencing chronic blood loss from the esophagus/stomach or small bowel, or a combination of both. EVs and PHG are reported to hemorrhage at an estimated yearly rate of 5–15% and 8–25%, respectively.28,29 No reports have been published, as far as we are aware, that have stated the estimated yearly rate of hemorrhage in patients with PHE. Currently, small-bowel angioectasias and varices are thought to account for the unexplained blood losses in patients with PHT,4,6 and some of these could be suitable candidates for endoscopic treatment.15 Furthermore, we should remember that even villous edema appears to be susceptible to bleeding.7 When the PHE lesions were categorized into two groups in this study in accordance with De Palma et al.,2 inflammatory-like abnormalities predominated among our patients (n = 84, 63%), Journal of Gastroenterology and Hepatology 30 (2015) 124–130
© 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
T Aoyama et al.
followed by vascular lesions (n = 33, 25%). When these lesions were evaluated based on the presence of PSs, the PHE lesions that were associated with PSs were inflammatory-like abnormalities, including erythema and villous edema, rather than vascular lesions. We did not assess the hepatic inflammatory markers, AST and ALT, as predictors of these lesions. Congestion in the portal system is considered the main mucosal alteration in the small bowel that occurs in a similar manner as PHG.30,31 Although the development of vascular lesions is considered similar to that of PSs, we could not demonstrate a close relationship between PSs and vascular lesions, perhaps because of the multifactorial processes that underlie PHE development that may manifest themselves diversely. Similarly, GVs were not significantly related to the presence of PHE, probably because of the anatomical and hemodynamic diversity presented by GVs.32 In relation to whether inflammatory-like abnormalities are more dependent on portal pressure, Takahashi et al. showed that edema of the small bowel was strongly correlated with portal venous pressure.33 Furthermore, Matsushita et al. reported that reddened lesions and edema in the small bowel were attenuated once the portal pressure was reduced by a transjugular intrahepatic portosystemic shunt, but that the number of angioectasias did not change because of the reduction in portal pressure.34 Further debate is required regarding the treatment of PHE lesions. Our study has some limitations. First, we did not compare our findings with those from a control healthy group because of the retrospective nature of the study. Second, the study included patients from a single center only. Therefore, a large-scale study is needed to address these limitations. In relation to the consequences of PHT, PSs appear to reliably predict PHE and we can conclude that PHE presence can be predicted. Moreover, CE should be considered for patients with LC accompanied by PSs, especially where there are left gastric vein and splenorenal shunts.
References 1 Martínez-Noguera A, Montserrat E, Torrubia S, Villalba J. Doppler in hepatic cirrhosis and chronic hepatitis. Semin. Ultrasound CT MR 2002; 23: 19–36. 2 De Palma GD, Rega M, Masone S et al. Mucosal abnormalities of the small bowel in patients with cirrhosis and portal hypertension: a capsule endoscopy study. Gastrointest. Endosc. 2005; 62: 529–34. 3 Viggiano TR, Gostout CJ. Portal hypertensive intestinal vasculopathy: a review of the clinical, endoscopic, and histopathologic features. Am. J. Gastroenterol. 1992; 87: 944–54. 4 Akyuz F, Pinarbasi B, Ermis F et al. Is portal hypertensive enteropathy an important additional cause of blood loss in portal hypertensive patients? Scand. J. Gastroenterol. 2010; 45: 1497–502. 5 Yamada A, Watabe H, Obi S et al. Surveillance of small intestinal abnormalities in patients with hepatocellular carcinoma: a prospective capsule endoscopy study. Dig. Endosc. 2011; 23: 124–9. 6 Tang SJ, Zanati S, Dubcenco E et al. Diagnosis of small-bowel varices by capsule endoscopy. Gastrointest. Endosc. 2004; 60: 129–35. 7 Higaki N, Matsui H, Imaoka H et al. Characteristic endoscopic features of portal hypertensive enteropathy. J. Gastroenterol. 2008; 43: 327–31.
Prediction of enteropathy
8 Raju GS, Gerson L, Das A, Lewis B. American Gastroenterological Association (AGA) Institute technical review on obscure gastrointestinal bleeding. Gastroenterology 2007; 133: 1697–717. 9 Kodama M, Uto H, Numata M et al. Endoscopic characterization of the small bowel in patients with portal hypertension evaluated by double balloon endoscopy. J. Gastroenterol. 2008; 43: 589–96. 10 Roberts-Thomson IC, Singh R, Teo E, Nguyen NQ, Lidums I. The future of endoscopy. J. Gastroenterol. Hepatol. 2010; 25: 1051–7. 11 Imagawa H, Oka S, Tanaka S et al. Improved visibility of lesions of the small intestine via capsule endoscopy with computed virtual chromoendoscopy. Gastrointest. Endosc. 2011; 73: 299–306. 12 Fukumoto A, Tanaka S, Shishido T, Takemura Y, Oka S, Chayama K. Comparison of detectability of small-bowel lesions between capsule endoscopy and double-balloon endoscopy for patients with suspected small-bowel disease. Gastrointest. Endosc. 2009; 69: 857–65. 13 Shishido T, Oka S, Tanaka S et al. Diagnostic yield of capsule endoscopy vs. double-balloon endoscopy for patients who have undergone total enteroscopy with obscure gastrointestinal bleeding. Hepatogastroenterology 2012; 59: 955–9. 14 Teshima CW, Kuipers EJ, van Zanten SV, Mensink PB. Double balloon enteroscopy and capsule endoscopy for obscure gastrointestinal bleeding: an updated meta-analysis. J. Gastroenterol. Hepatol. 2011; 26: 796–801. 15 Aoyama T, Oka S, Aikata H et al. Small bowel abnormalities in patients with compensated liver cirrhosis. Dig. Dis. Sci. 2013; 58: 1390–6. 16 Goulas S, Triantafyllidou K, Karagiannis S et al. Capsule endoscopy in the investigation of patients with portal hypertension and anemia. Can. J. Gastroenterol. 2008; 22: 469–74. 17 Jeon SR, Kim JO, Kim JB et al. Portal hypertensive enteropathy diagnosed by capsule endoscopy in cirrhotic patients: a nationwide multicenter study. Dig. Dis. Sci. 2014; 59: 1036–41. 18 Trevisani F, De Notariis S, Rapaccini G et al. Semiannual and annual surveillance of cirrhotic patients for hepatocellular carcinoma: effects on cancer stage and patient survival (Italian experience). Am. J. Gastroenterol. 2002; 97: 734–44. 19 El-Serag HB, Marrero JA, Rudolph L, Reddy KR. Diagnosis and treatment of hepatocellular carcinoma. Gastroenterology 2008; 134: 1752–63. 20 Weishaupt D, Pfammatter T, Hilfiker PR, Wolfensberger U, Marincek B. Detecting bleeding duodenal varices with multislice helical CT. AJR Am. J. Roentgenol. 2002; 178: 399–401. 21 Li X, Chen H, Dai J, Gao Y, Ge Z. Predictive role of capsule endoscopy on the insertion route of double-balloon enteroscopy. Endoscopy 2009; 41: 762–6. 22 Zardi EM, Uwechie V, Caccavo D et al. Portosystemic shunts in a large cohort of patients with liver cirrhosis: detection rate and clinical relevance. J. Gastroenterol. 2009; 44: 76–83. 23 Nagaoki Y, Aikata H, Kobayashi T et al. Risk factors for the exacerbation of esophageal varices or portosystemic encephalopathy after sustained virological response with IFN therapy for HCV-related compensated cirrhosis. J. Gastroenterol. 2013; 48: 847–55. 24 Hino S, Kakutani H, Ikeda K et al. Hemodynamic assessment of the left gastric vein in patients with esophageal varices with color Doppler EUS: factors affecting development of esophageal varices. Gastrointest. Endosc. 2002; 55: 512–7. 25 Bosch J, Abraldes JG, Berzigotti A, Garcia-Pagan JC. Portal hypertension and gastrointestinal bleeding. Semin. Liver Dis. 2008; 28: 3–25. 26 Kumamoto M, Toyonaga A, Inoue H et al. Long-term results of balloon-occluded retrograde transvenous obliteration for gastric fundal varices: hepatic deterioration links to portosystemic shunt syndrome. J. Gastroenterol. Hepatol. 2010; 25: 1129–35.
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
129
Prediction of enteropathy
T Aoyama et al.
27 Barakat M, Mostafa M, Mahran Z, Soliman AG. Portal hypertensive duodenopathy: clinical, endoscopic, and histopathologic profiles. Am. J. Gastroenterol. 2007; 102: 2793–802. 28 North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices. Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices. A prospective multicenter study. N. Engl. J. Med. 1988; 319: 983–9. 29 D’Amico G, Pagliaro L, Bosch J. The treatment of portal hypertension: a meta-analytic review. Hepatology 1995; 22: 332–54. 30 Toyoda H, Nakao M, Ogura Y et al. Congestion of superior mesenteric veins and small bowel mucosal injury after endoscopic treatment of esophageal varices in patients with portal hypertension. Dig. Dis. Sci. 2001; 46: 2353–9.
130
31 Misra V, Misra SP, Dwivedi M, Gupta SC. Histomorphometric study of portal hypertensive enteropathy. Am. J. Clin. Pathol. 1997; 108: 652–7. 32 Saad WE. Vascular anatomy and the morphologic and hemodynamic classifications of gastric varices and spontaneous portosystemic shunts relevant to the BRTO procedure. Tech. Vasc. Interv. Radiol. 2013; 16: 60–100. 33 Takahashi Y, Fujimori S, Narahara Y et al. Small intestinal edema had the strongest correlation with portal venous pressure amongst capsule endoscopy findings. Digestion 2012; 86: 48–54. 34 Matsushita Y, Narahara Y, Fujimori S et al. Effects of transjugular intrahepatic portosystemic shunt on changes in the small bowel mucosa of cirrhotic patients with portal hypertension. J. Gastroenterol. 2013; 48: 633–9.
Journal of Gastroenterology and Hepatology 30 (2015) 124–130 © 2014 Journal of Gastroenterology and Hepatology Foundation and Wiley Publishing Asia Pty Ltd
