A total of 152 consecutive children with oesophageal varices have been endoscopically reviewed since 1979. In all, 108 of these children presented with variceal bleeding which was managed by injection sclerotherapy. Variceal obliteration was achieved in 33 (92 per cent) children with extrahepatic portal hypertension and 54 (75 per cent) with intrahepatic portal hypertension. Prophylactic injection sclerotherapy was used to obliterate large varices in 11 children with no history of haemorrhage. Bleeding episodes occurred in 38 (39 per cent) children before variceal obliteration was complete. However, the mortality rate from variceal bleeding was only 1 per cent. Complications were oesophageal ulceration (29 per cent) and stricture (16 per cent) which both resolved with conservative management. During a mean follow-up period of 2.9 years after sclerotherapy, recurrent oesophageal or gastric varices developed in 12 (12 per cent) cases, with rebleeding in 9 (9 per cent), but all responded successfully to a second course of treatment. These results are superior to contemporary surgical management and injection sclerotherapy should therefore currently be the primary treatment of choice for bleeding oesophageal varices in children.

The incidence of venous thromboembolism after elective knee surgery has previously been studied almost exclusively in patients receiving total knee replacements, in whom the risk of a deep vein thrombosis is approximately 60%. We report the results of ipsilateral ascending venography in 312 patients undergoing a wide variety of elective knee operations under tourniquet ischaemia, none of whom received any specific prophylaxis against thromboembolism. Total knee replacement was confirmed to carry a high risk with ipsilateral deep vein thrombosis in 56.4% and symptomatic pulmonary embolism in 1.9%. By contrast, arthroscopy was associated with a low incidence of venous thrombosis (4.2%). Meniscectomy, arthrotomy, patellectomy, synovectomy and arthrodesis were all high-risk procedures, particularly in patients over 40 years of age, and were associated with deep vein thrombosis rates of 25% to 67%. On the basis of these findings, we advise prophylaxis against venous thromboembolism in all patients over 40 years of age undergoing elective knee surgery other than arthroscopy.

Cystic dilatation of the biliary tree is a rare congenital anomaly. To determine mode of presentation, diagnostic pitfalls, and long term outcome after surgery, 78 children (57 girls, 21 boys) with choledochal cyst treated between 1974 and 1994 were reviewed. Anatomical types were: Ic (n=44), If (n=28), IVa (n=4), and V (n=2); a common pancreaticobiliary channel was identified in 76% patients. Age at presentation ranged from 0-16 (median 2.2) years, six patients being diagnosed by prenatal ultrasonography. Of the 72 patients diagnosed postnatally, 50 (69%) presented with jaundice, associated with abdominal pain in 25 or a palpable mass in three, 13 (18%) presented with pain alone, and two (30/o) with a palpable mass. The classic triad of jaundice, pain, and a right hypochondrial mass was present in only four (6%). Four children presented acutely after spontaneous perforation of a choledochal cyst, two presented with ascites and one cyst was discovered incidentally. Plasma and/or biliary amylase values were raised in 30 of 31 patients investigated for abdominal pain; seven had evidence of pancreatitis at operation. In 35 of 67 (52%) patients referred without previous surgery, symptoms had been present for more than one month, and in 14 of them for more than one year, before diagnosis. Delayed referral was due to misdiagnosis as hepatitis (n= 12), incomplete investigation of abdominal pain (n=6), and failure to note the significance of ultrasonographic findings (n= 10). Two patients referred late died from liver failure. Of the 76 patients with type I or IV cysts, 59 underwent radical cyst excision and hepaticojejunostomy as a primary procedure and 10 as a secondary operation after previously unsuccessful surgery. Sixteen patients have been lost to follow up but most of the remainder are well after a mean period of 4-1 (0.1-13) years. Choledochal cysts are often misdiagnosed, but prognosis is excellent if radical excision is performed. (Arch Dis Child 1995; 73: 528-531)

The results of surgery in 48 patients with hepatic hydatid disease are described. In 26 out of 32 patients with uncomplicated cysts, conservative surgery with obliteration of the cyst cavity by omentoplasty gave satisfactory results. In contrast, 16 patients with cysts complicated by pyogenic infection or cholangitis required a variety of surgical procedures, including prolonged tube drainage, exploration of the common bile duct, sphincterotomy/plasty and hepatojejunostomy to achieve a satisfactory outcome. The group with complicated cysts also had a much longer postoperative stay (median 30 days) than those with uncomplicated cysts (median 14 days). Two patients died: in one, post-mortem examination revealed carcinoma in the wall of the cyst.

Plasma lipid peroxide concentrations were measured in 100 patients with occlusive arterial disease proved angiographically (50 patients with ischaemic heart disease, 50 with peripheral arterial disease) and compared with values in 75 control patients with no clinical evidence of atherosclerosis. Lipid peroxide concentrations were significantly higher in patients both with ischaemic heart disease (median 4-37 iimol/l (interquartile range 3-85-5-75 [tmol/l); p<0-001) and with peripheral arterial disease (median 4-37 [smoltl (3-88-5-21 [mol/l); p<0001) than in controls (median 3-65 [tmol/l (interquartile range 3-29-3-89 [mol/l)).Overall there was a significant but weak correlation between plasma lipid peroxide and plasma triglyceride concentrations (r,=0-25; p<0-001) but not between plasma lipid peroxide and plasma total cholesterol concentrations. Furthermore, hypertension, obesity, diabetes, smoking, positive family history, and intake of fi blockers and thiazide diuretics were not associated with significant differences in lipid peroxide values.This study provides clinical support to experimental data indicating that peroxidised lipids may be important in atherogenesis and its complications and also suggests that peroxidised lipids may provide an index of the severity of atherosclerosis.Thrombosis Research

Aim-To quantify the exposure of very low birthweight neonates to ionising radiation from diagnostic x-rays. Methods-Retrospective analysis was made of all radiographs performed over 18 months in an integrated special care baby unit and regional neonatal surgical unit in a large teaching hospital of surviving inborn babies of very low birthweight (<1500 g) admitted to the unit. Results-Fifty five VLBW neonates were treated for a total of 3296 days and received 498 radiographs (median 5 per infant). The mean eVective radiation dose was 0.04 mSv and the maximum for one infant was 0.54 mSv. Conclusions-The radiation burden in this group of neonates is low and the benefits of diagnostic radiographs far outweigh any potential radiation risks. (Arch Dis Child Fetal Neonatal Ed 1998;78:F227-F229)

No abstract

Clinical and angiographic findings in 53 patients with portal vein occlusion were reviewed to determine the relationship between the pattern of venous occlusion and its aetiology. Five major patterns were identified. Group 1 comprised three children with idiopathic intrahepatic portal vein occlusion alone (hepatoportal sclerosis). Those in group 2 (25 patients), with occlusion of the main portal vein, accounted for almost half the total number and for most of those with a probable congenital aetiology. The ten patients in group 3 had angiographic occlusion and collateralization of the main portal and superior mesenteric veins. Intra-abdominal sepsis, other than that from pancreatic disease, was associated with this pattern of venous occlusion in which the splenic vein is spared. No patients with isolated portal and splenic vein occlusion (group 4) were recorded. Widespread thrombosis throughout the portal venous system (group 5; 15 patients) had a multifactorial aetiology but was typically observed in hypercoagulable states that were also associated with hepatic vein occlusion or other deep vein thromboses.