Gut, 1991,32,309-311 309

Anorectal varices - their frequency in cirrhotic and

non-cirrhotic portal hypertension Gut: first published as 10.1136/gut.32.3.309 on 1 March 1991. Downloaded from

Y Chawla, J B Dilawari

Abstract group included 37 patients with extrahepatic Anorectal varices in portal hypertension have portal venous obstruction (mean (SD) age 18.8 been little studied. Seventy eight per cent of (6.8) years) diagnosed by splenoportoveno- 72 patients with portal hypertension had graphy and 10 patients with non-cirrhotic portal anorectal varices shown at flexible sigmoido- fibrosis (mean (SD) age 28.6 (9.6) years) diag- scopy. Significantly more patients with non- nosed by splenoportovenography and liver cirrhotic portal hypertension had these varices biopsy according to the criteria laid down by than patients with cirrhosis (89% v 56%, Indian Council of Medical Research.3 All these pbleeding. Fifty five (76%) of the 72 patients had Portal hypertension leads to the development of presented with an upper gastrointestinal bleed in collaterals between the portal and systemic cir- the past and were thus put on a sclerotherapy culations, namely at the gastroesophageal junc- programme. Only one patient presented with tion, anal region, falciform ligament, and areas bleeding per rectum. where the abdominal organs are in contact with All patients were assessed for oesophageal retroperitoneal tissues. These collateral vessels varices with an upper gastrointestinal endoscope may not develop to the same degree since the (Olympus GIF X Q) and anorectal or colonic amount of blood flowing through them may varices (limited to the rectum, sigmoid, and differ. Because ofthis, varices at one region may descending colon) with a flexible sigmoidoscope be small and those at another large. Hosking et al (Olympus CF, PIOS). Colonic preparation for recently showed a frequency of anorectal varices sigmoidoscopy was done by giving 200 ml of20% of 44% in their group of cirrhotic patients.' We mannitol orally three to four hours before the have previously reported the frequency of sigmoidoscopic examination. Flexible sig- http://gut.bmj.com/ anorectal varices in a small number of patients moidoscopy was undertaken once the patient with portal hypertension from our centre.2 Since developed loose clear stools. we have an appreciable proportion of patients Anorectal varices were diagnosed if bluish or with non-cirrhotic portal hypertension we inves- grey distended tortuous or sacular veins were tigated the frequency of anorectal varices in this seen above the anal margin and extending into group ofpatients and compared it with the value the rectum. They were described as small or found in patients with cirrhosis. large if their diameter was less or more than on September 29, 2021 by guest. Protected copyright. 5 mm respectively. Similarly, oesophageal varices were said to be small (grade I-II) or large Patients and methods (grade III-IV) if their diameter was less or more We studied 72 consecutive patients with portal than 5 mm respectively. hypertension. Twenty five patients had cirrhosis The anorectal varices were shown to a second proved at biopsy (mean (SD) age 44.5 (10.8) observer and an independent assessment was years) and 47 had non-cirrhotic portal hyperten- made. sion (mean (SD) age 20.9 (8.44) years). The aetiology of cirrhosis in the 25 patients with this disorder was alcohol in 16, hepatitis B virus in Results four, and cryptogenic in five. The non-cirrhotic Anorectal varices were observed in 56 (77.7%) of

TABLE I Frequeny (%) ofanorectal varices, oesophageal varices, and uppergastrointestinal (UGI) bleeding in patients with non-cirrhotic portal hypertension (NCPH) and cirrhosis (C) Non-cirrhotic patients Statistical significance (p) Cirrhotic Department of EHPO NCPF NCPH patients NCPH v EHPO v NCPFv NCPFv Hepatology, (n=37) (n=10) (n=47) (n=25) C C EHPO C Postgraduate Institute of Anorectal varices: Medical Education and Large 16(43) 4(40) 20(42) 4(16) <0 05 <0 05 NS NS Research, Chandigarh, Small 19(51) 3 (30) 22 (47) 10(40) NS NS NS NS India Large+small 35 (94) 7 (70) 42(89) 14(56) <0-01 <0-01 NS NS Y Chawla Absent 2 (5) 3 (30) 5 (10) 11(44) <0-01 <0-01 NS NS J B Dilawari Oesophageal varices: Large 34 (92) 9 (90) 43 (91) 17(68) <0 05 <0 05 NS NS Correspondence to: Small 3 (8) 1(10) 4(8) 6 (24) NS NS NS NS Dr J B Dilawari, Department Large+small 37 (100) 10(100) 47 (100) 23 (92) NS NS NS NS of Hepatology, Postgraduate Absent 0 0 0 2 (8) NS NS NS NS Institute of Medical UGI bleed: Education and Research, Present 35 (95) 7 (70) 42 (89) 13 (52) <0-01 <0-01 NS NS Chandigarh-160 012, India. Absent 2 (5) 3 (30) 5 (10) 12 (48) <0-01 <0-01 NS NS Accepted for publication 12 March 1990 EHPO=extrahepatic portal venous obstruction; NCPF=non-cirrhotic portal fibrosis; NS=not significant. 310 Chawla, Dilawari

TABLE II Correlation ofanorectal varices with upper by Hosking etal.' This difference is related to the gastrointestinal (GI) bleeding different population of patients studied in the two series. Two thirds of our patients were non- UpperGI bleed (no (%)) Gut: first published as 10.1136/gut.32.3.309 on 1 March 1991. Downloaded from Anorectal cirrhotic and a third were cirrhotic whereas all varices Present Absent p their patients had cirrhosis.' Our cirrhotic Large 21(38) 3 (17-6) NS patients had similar frequency of anorectal Small 26(38) 6 (35.2) NS Large+small 47 (85 4) 9 (53) <0-02 varices to that reported by Hosking et al. l Absent 8 (14-5) 8 (47) <0-02 Indeed, we found a significantly higher incidence of varices in patients with non- cirrhotic portal hypertension (especially extra- the 72 patients with portal hypertension. Forty hepatic portal venous obstruction) than in two (89 3%) ofthe 47 patients with non-cirrhotic cirrhotic patients. This is not surprising since portal hyptertension had anorectal varices com- patients with non-cirrhotic portal hypertension, pared with 14 (56%) of the 25 patients with present with complications of portal hyperten- cirrhosis (p<0-01). Anorectal varices were also sion alone, which are much greater and of longer significantly more common in extrahepatic standing - for example upper gastrointestinal portal venous obstruction compared with bleed - than those found in patients with cirrho- cirrhosis (p<0-01) (Table I). Large anorectal sis, where other complications of liver disease varices were seen in 24 patients (16 extrahepatic such as ascites and encephalopathy are likely to portal venous obstruction, four non-cirrhotic precipitate a presentation. Though patients with portal fibrosis, and four cirrhosis) and small in 32 non-cirrhotic portal hypertension differed patients (19 extrahepatic portal venous obstruc- significantly from patients with cirrhosis in tion, three non-cirrhotic portal fibrosis, 10 terms of presence of anorectal varices, oesopha- cirrhosis). Significantly more patients with non- geal varices, and upper gastrointestinal bleeding cirrhotic portal hypertension and extrahepatic as shown in Table I, this difference was mainly portal venous obstruction had large anorectal contributed by patients with extrahepatic portal varices compared with cirrhosis patients venous obstruction. However, the number of (p<005). Upper gastrointestinal bleeding as a patients with non-cirrhotic portal hypertension presenting feature was also significantly more in this study was not great enough to make a common in non-cirrhotic portal hypertension categorical statement on whether these patients and extrahepatic portal venous obstruction com- behave more like those with extrahepatic portal pared with cirrhosis (Table I). Ofthe 55 patients venous obstruction or cirrhosis. It has also been who presented with an upper gastrointestinal observed by other authors that ectopic varices

bleed, 47 (85.4%) had evidence of anorectal are more commonly observed in patients with http://gut.bmj.com/ varices, which was significantly higher than in portal venous obstruction than in those with patients without bleeding (p<002) (Table II). cirrhosis.7 Oesophageal varices were detected in 70 (97%) Unlike oesophageal varices, anorectal varices of the 72 patients (Table I). Large oesophageal rarely bleed. Bleeding anorectal varices have varices were significantly more common in been reported by Wilson et al in two of 309 patients with non-cirrhotic portal hypertension patients with oesophageal varices,' while McCor-

and extrahepatic portal venous obstruction than mack observed this in four of 112 patients with on September 29, 2021 by guest. Protected copyright. in cirrhotic patients. Although more patients portal hypertension.9 In a larger series by John- with large oesophageal varices had large son, rectal varices were reported in five of their anorectal varices compared with those patients 1100 patients,'0 while Hosking et al observed with small oesophageal varices, the difference bleeding anorectal varices in only two of 100 was not statistically significant. patients with portal hypertension.' In our series, There was no significant difference in the only one patient with extrahepatic portal venous frequency of anorectal varices, oesophageal obstruction bled from anorectal varices. varices, and upper gastrointestinal bleeding Our findings agree with those of Hosking et al when patients with non-cirrhotic portal fibrosis that anorectal varices reflect a later stage in the were compared with extrahepatic portal venous development of portal hypertension.' This is obstruction and cirrhosis. supported by the fact that patients with large None of the patients studied had evidence of oesophageal varices have more frequent ano- varices in the sigmoid or descending colon. rectal varices than patients with small or no oesophageal varices. Moreover, the incidence of anorectal varices was significantly higher in Discussion patients who had an upper gastrointestinal Portal hypertension in most patients results in bleed, which, as a complication, occurs later in the development of oesophagogastric varices the course ofportal hypertension. which are associated with massive upper gastro- Anorectal varices are best shown by endo- intestinal bleeds. Rectal varices, on the other scopy, since it is a direct means of establishing a hand, constitute another collateral pathway, diagnosis. We preferred to use a flexible sig- which helps in decompressing the portal system moidoscope in assessing anorectal varices, since into the systemic circulation through the patients tolerate it better and the bowel lumen superior middle and inferior haemorrhoidal can be distended with air, allowing a better view veins." They are not usually associated with of the rectum than with rigid sigmoidoscope." appreciable morbidity. We believe that the flexible sigmoidoscope is as Our finding of anorectal varices in 78% of sensitive as the proctoscope in showing anorectal patients with portal hypertension is a signifi- varices, although internal piles that are not cantly higher percentage than the 44% reported anorectal varices may be missed. 12 Anorectal varices - theirfrequency in cirrhotic and non-cirrhotic portal hypertension 311

The incidence of colonic varices in portal 3 Workshop on non-cirrhotic portal fibrosis. Indian Council of Medical Research, New Delhi: 1969: 36. hypertension has not been reported previously. 4 Edwards FA. Functional anatomy of portosystemic com- Only 24 cases of bleeding colonic varices from munications. Arch Intern Med 1951: 88:137-54. 5 Britton RC. Influence ofporto-systemic collateral patterns and Gut: first published as 10.1136/gut.32.3.309 on 1 March 1991. Downloaded from the sigmoid or descending colon have been distribution of varices on results of surgical treatment of reported.7 We did not find sigmoid or descend- bleeding esophageal varices Surgery 1963; 53: 567-74. 6 Fleming RJ, Seeman WB. Roentgenographic demonstration ing colon varices in any of the 72 patients we of unusual extra esophageal varices. AmJ Roentgenol 1968; studied. 103: 281-90. 7 Lebrec D, Benhamau JP. Ectopic varices in portal hyperten- In conclusion, in our series significantly more sion. Clin Gastroenterol 1985; 14: 105-21. patients with non-cirrhotic portal hypertension 8 Wilson SE, Stone RT, Christie JP, Panaro E. Massive lower gastrointestinal bleeding from intestinal varices. Arch Surg had anorectal varices, compared with patients- 1979; 114: 1158-61. with cirrhosis. Even if the varices are large, 9 McCormack TT, Barkley HR, Simms JM, Johnson AG. Rectal varices are not piles. BrJ Surg 1984; 74: 163. however, they do not usually bleed. 10 Johnson K, Bardin J, Orloff MJ. Massive bleeding from hemorrhoidal varices in portal hypertension. JAMA 1980; 244: 2084-5. 11 Bohlman TW, Katon RM, Lipshutz GR, McCool MF, 1 Hosking SW, Johnson AG, Smart HL, Triger DR. Anorectal Smith FW, Clifford SM. Fibroptic pansigmoidoscopy. varices, haemorrhoids and portal hypertension. Lancet 1989; Gastroenterology 1977; 72: 644-9. i: 349-52. 12 Bernstein WC. What are hemorrhoids and what is their 2 Chawla YK, Dilawari JB. Anorectal varices haemorrhoids and relationship to the portal system? Dis Colon Rectum 1983; 26: portal hypertension. Lancet 1989; i: 725. 829-34. http://gut.bmj.com/ on September 29, 2021 by guest. Protected copyright.