- Email: [email protected]
Cost-effectiveness of N-butyl-2-cyanoacrylate (histoacryl) glue injections versus transjugular intrahepatic portosystemic shunt in the management of acute gastric variceal bleeding
THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2003 by Am. Coll. of Gastroenterology Published by Elsevier Inc.
Vol. 98, No. 12, 2003 ISSN 0002-9270/03/$30.00 doi:10.1016/j.amjgastroenterol.2003.09.039
Cost-Effectiveness of N-Butyl-2-Cyanoacrylate (Histoacryl) Glue Injections Versus Transjugular Intrahepatic Portosystemic Shunt in the Management of Acute Gastric Variceal Bleeding S. Mahadeva, M.B., M.R.C.P., M. C. Bellamy, M.A., M.B., F.R.C.A., D. Kessel, M.A., M.R.C.P., F.R.C.R., M. H. Davies, M.D., F.R.C.P., and C. E. Millson, M.D., F.R.C.P. Departments of Hepatology and Liver Transplantation, Intensive Care Medicine, and Radiology, St. James’s University Hospital, Leeds, United Kingdom
ABSTRACT The management of bleeding gastric varices has not been standardized. Although transjugular intrahepatic portosystemic shunt (TIPS) is used in most centers, endoscopic treatment with N-butyl-2-cyanoacrylate (cyanoacrylate) glue has recently been shown to be effective. Cost-effectiveness analyses of these methods are lacking. METHODS: We performed a retrospective review of patients with bleeding gastric varices treated either by TIPS or cyanoacrylate glue injection. Economic analysis was based on direct costs for a fixed financial year. The two groups were compared for a period of 6 months follow-up, to liver transplantation, or death for each patient. RESULTS: Between January, 1995 and December, 1999, 20 patients with bleeding gastric varices had TIPS; 23 patients had cyanoacrylate glue injection from January, 2000 to October, 2001. There were no significant differences between the two groups in patient characteristics, transfusion requirement, and gastric variceal anatomy. In the TIPS group, 15/20 patients had the procedure performed within 24 h of hemorrhage, and 90% of stent insertions were successful. Complications consisted of two cases of pulmonary edema, two cases of severe encephalopathy, and a 15% stenosis rate at 6 months. In the glue group, there were 3 ⫾ 1.5 endoscopies and 2 ⫾ 1 injections per patient, with a 96% initial hemostasis. There was one case of (glue) pulmonary embolism and one blocked front endoscope lens, which required repair. The initial rebleed rate was significantly lower in patients who had TIPS (15% vs 30%, p ⫽ 0.005). The inpatient stay was shorter in the glue group (13 ⫾ 1 vs 18 ⫾ 2 days, p ⫽ 0.05), but there was no difference in the overall mortality rate. The median cost within 6 months of initial gastric variceal bleeding was $4,138 ($3,009 –$8,290) for glue versus $11,906 ($8,200 –$16,770) for TIPS (p ⬍ 0.0001). CONCLUSION: In this comparable group of patients, cyanoacrylate glue injection was more cost effective than TIPS in
the management of acute gastric variceal bleeding. A prospective, randomized trial would be required to confirm our analysis. (Am J Gastroenterol 2003;98:2688⫺2693. © 2003 by Am. Coll. of Gastroenterology)
INTRODUCTION The optimal management of bleeding gastric varices remains uncertain. Standard endoscopic techniques do not seem to be effective (1–3), and surgical shunting is associated with a high mortality (4, 5); hence, transjugular intrahepatic portosystemic shunt (TIPS) has emerged as the recommended treatment. Although TIPS achieves hemostasis in more than 90% of bleeding gastric varices (6, 7), the early mortality rate for emergently placed TIPS in patients with advanced liver disease has been reported to be between 40% and 60% (8, 9), and recurrent bleeding ranges from 29% (6) to 53% (10). Recently, endoscopic treatment with cyanoacrylate glue has been reported as an effective option in the management of bleeding gastric varices (11, 12). Large series from Asia and Europe have shown primary hemostasis rates of 94 – 97%, low rebleeding rates, and early mortality rates of only 12–20% (13–15), prompting several authors to propose cyanoacrylate glue as the first choice for bleeding gastric varices (11–13). Because neither TIPS nor endoscopic cyanoacrylate injection actually treats the underlying cause of the gastric varices, the effectiveness and economic impact of the technique should influence the clinician’s choice of therapy. It has been suggested that endoscopic treatment with cyanoacrylate is probably cheaper than TIPS (12), but this has not been proved to date. The aim of this study was to compare, retrospectively, the outcomes and health care costs of patients treated with either cyanoacrylate glue or TIPS as the primary treatment. Costs were evaluated with an intent-to-treat analysis and included rescue therapies for both modalities in the evaluation. Out-
AJG – December, 2003
come measures included rebleed rates, transfusion requirement, and survival.
Glue vs. TIPS for Bleeding Gastric Varices
Table 1. Institutional Charges as a Measure of Direct Costs of the Parameters Analyzed Cost Items
MATERIALS AND METHODS All patients referred to this institution for bleeding gastric varices from 1995 were identified from our TIPS database (department of vascular radiology) and endoscopy unit (Endoscribe database, Astra, Endoscribe, Medlboss, Adelaide, Australia). Only patients with confirmed bleeding gastric varices on upper GI endoscopy were included in the study. This was defined as evidence of visible or active bleeding at the time of endoscopy, the presence of fresh clot on varices, or the presence of “red spots” on gastric varices with fresh or altered blood in the stomach without any other possible source of bleeding. Relevant clinical data were reviewed from patients’ clinical records, and endoscopic data were recorded from clinical notes and computerized endoscopy reports. TIPS had been performed under general anesthesia with a standard technique described previously (16). Only Wallstents (Pfeizer, Bulach, Switzerland) or Memotherm (Angiomed, Karlsruhe, Germany) stents had been used. After insertion, routine Doppler ultrasound scanning would be performed after 2 days and after 1 wk and then on an every-3-months basis to assess stent patency. If shunt dysfunction was suspected on Doppler scan, angiography would be performed. The technique of endoscopic cyanoacrylate glue injection used in our unit has been described previously (13). At endoscopy, N-butyl-2-cyanoacrylate (Histoacryl, B. Braun, Melsungen, Germany) was diluted with Lipiodol (Laboratoire Guerbet, Aulnay-Sousels, France) in a 1:1.5 ratio and injected as a bolus of 1 to 2 ml, according to variceal size. Most patients had a plain abdominal x-ray postendoscopy to evaluate opacification of the varices. Follow-up post–index endoscopy was arranged within 48 h, then on a weekly or monthly basis, depending on the degree of variceal obliteration. The latter was assessed by inspection and blunt prodding of the variceal mass at endoscopy. Because the patients who had been treated with cyanoacrylate had a shorter follow-up period (median of 6 months) than those originally treated with TIPS, cost comparison between the two groups of patients were limited to 6 months post–index bleed, time of liver transplantation, or death. During the period studied, days of hospitalization owing to the initial index bleeding episode and further hospitalizations for rebleeding or complications of treatment were recorded as the cost of the therapeutic procedures. The cost of TIPS included all equipment, time of medical and radiologic staff, medication, and allotted 2 h for general anesthesia. Similarly, the cost of endoscopic cyanoacrylate injection involved all equipment, the time of medical and nursing staff, and use of the endoscopy unit sessions. Estimates of inpatient stay unit costs include nursing staff costs, administrative and clerical staff costs, consumables, equipment, overheads, and capital costs. We assumed that
2689
Inpatient stay (per/day) Liver ward General medical ward Intensive care unit Radiology TIPS TIPS revision TIPS venogram Ultrasound scan ⫹ Doppler Endoscopy Upper GI endoscopy Upper GI endoscopy ⫹ glue (5 injections) Surgery Mesocaval shunt Splenectomy and gastric devascularisation
US$ 195.2 101.9 2159 5080 2030 610.16 102.45 150.80 389.38 8027 6291
there would be no differences in ward staff fee, routine blood investigations, standard vasoactive drugs, and basic radiology (i.e., chest x-ray) between the two groups because they had been managed in the same hospital with little difference in practice apart from the principle therapeutic modality. All costs were based on the Healthcare Resource Grouping codes (National Health Service, United Kingdom) of this institution over the April, 2000 to March, 2001 financial period (Table 1). Cost analysis of each therapeutic modality was calculated on an intent-to-treat basis. Survival and rebleeding were calculated according to the Kaplan-Meier method, and differences between treatment groups were analyzed by the log-rank test. Quantitative and qualitative variables were compared with either MannWhitney U test or Fisher exact test, where appropriate. The cost difference for the period of study was analyzed with the Mann-Whitney U test. Statistical significance was assumed at p ⬍ 0.05.
RESULTS Between January, 1995 and December, 1999, 20 patients underwent TIPS insertion for confirmed bleeding gastric varices. Between January, 2000 and October, 2001, a further 23 patients with bleeding gastric varices were treated with cyanoacrylate glue. There were no significant differences between the two treatment groups in terms of patient demographics, etiology of liver disease, Child-Pugh classification, and gastric variceal anatomy (Table 2). The latter had been based on the classification system used by Sarin et al. (17). To assess the severity of bleeding in both groups, hemodynamic and hematological parameters were compared (Table 3). There was no significant hemodynamic difference between the two groups of patients on the day of admission during the index bleed, but there was a higher mean serum Hb (8.7 ⫾ 0.46 vs 7.6 ⫾ 0.35 g/dl, p ⫽ 0.038) and a slightly lower transfusion requirement (8 ⫾ 1 vs 9 ⫾ 2 U; mean difference 2.3, 95% CI ⫽ ⫺1.8 to 6.4) in the
2690
Mahadeva et al.
AJG – Vol. 98, No. 12, 2003
Table 2. Characteristics of Patients in the Study
Age (yr) Sex (male/female) Etiology Alcohol Cholestatic Cryptogenic Others Child-Pugh class A B C GV anatomy GOV 1 GOV 2
TIPS (n ⫽ 20)
Cyanoacrylate Glue (n ⫽ 23)
p
52 ⫾ 3 13/7
55 ⫾ 3 15/8
ns* ns†
12 3 2 3
14 3 3 3
ns†
4 6 10
4 8 9
ns†
13 7
15 8
ns†
GOV ⫽ gastroesophageal varices; GV ⫽ gastric varices. * Mann-Whitney U test. † Fisher exact test.
group treated with cyanoacrylate. At the index endoscopy, active bleeding was present in nine patients who received cyanoacrylate glue, compared with 12 patients in the TIPS group; a fresh adherent clot was seen in six patients in the glue group, compared with eight patients who had TIPS (p ⫽ ns). In the TIPS group, 15/20 patients had the procedure within 24 h of hemorrhage. Endoscopic treatment (two with thrombin, two with sodium tetradecyl sulfate) was unsuccessfully attempted in a further four patients before stent insertion. Eleven patients required either a SengstakenBlakemore or Minnesota tube for stabilization as well. Out of 20 stent insertions, 18 (90%) were successful. In the remaining two, shunting was unsuccessful despite two stent deployments in each of both cases; one patient rebled and eventually required a mesocaval shunt. The mean inferior Table 3. Hematological and Hemodynamic Parameters in Patients (Mean ⫾ SEM)
Hemoglobin (g/dL)† Heart rate† Systolic blood pressure (mm Hg)† Blood transfused (U) FFP transfused (U)
TIPS (n ⫽ 20)
Cyanoacrylate Glue (n ⫽ 23)
p*
7.6 ⫾ 0.35
8.7 ⫾ 0.46
0.038
90 ⫾ 4.2 113 ⫾ 7
94 ⫾ 3.4 117 ⫾ 4
ns ns
9⫾2
8⫾1
ns
6⫾1
5⫾1
ns
FFP ⫽ fresh frozen plasma. * By Mann-Whitney U test. † Measurements made on admission to ward.
vena cava–portal vein gradient dropped from 21 ⫾ 2 mm Hg to 5 ⫾ 0.7 mm Hg after TIPS insertion (p ⬍ 0.001). One patient continued to bleed despite shunting, and embolization of the varices was performed at the same time as TIPS. The median follow-up period among all cases was 12 months (range 0.1– 60). Early complications occurred in four cases: two with severe encephalopathy and two with pulmonary edema. All but one responded to medical therapy. The TIPS stenosis rate was 15% (three patients) over the 6-month cost analysis period; two presented with recurrent variceal bleeding, and the third was detected during routine surveillance. Further stent deployment was performed in two of these patients. Two patients had a liver transplant 8 days and 6 wk after the index bleed, respectively. Both had survived up to the follow-up period. In the group of patients that received cyanoacrylate glue, 10 patients required a Minnesota tube for stabilization. Out of 23 patients, 20 received cyanoacrylate glue injection within 24 h. Primary hemostasis was achieved in 96% of cases. The mean number of endoscopies with cyanoacrylate glue injection, to achieve variceal obliteration, was 2 ⫾ 1 (range 1–5) per patient. A mean of 7.5 ml (three vials) of cyanoacrylate was used during each endoscopic gluing session. Over the 6-month period, there were a mean of 3 ⫾ 1.5 (range 2–7) endoscopies performed per patient. The median follow-up period was 6 months (range 0.25–20 months). Post– glue injection abdominal or chest x-rays were performed in 17 patients (Table 4). Of these, seven had glue distribution that clearly appeared intravascular, and six appeared extravascular or intramural. Two had contrast in an intramural distribution but had varices outlined as well. There was no correlation between distribution of glue on x-ray and rebleed tendency. Ten had evidence of glue in the chest as well as the injection site in the gastric fundus. Whether this glue was transmitted by a vascular route or by aspiration of remnant glue/mucus mix from the stomach is uncertain. One patient with evidence of glue within the lung died of respiratory failure 24 h after the procedure. However, aspiration pneumonia is common in this group of patients, and glue was clearly visible in a significant percentage of patients receiving cyanoacrylate glue. A further complication consisted of a single case of occlusion of an endoscope (Olympus GIF IT240, Olympus, Southend-onSea, UK) front lens, which required repairing by the manufacturers. The cost of this was included in our analysis. Early (defined as less than 30 days) rebleeding occurred in 7/23 patients (30%) who had cyanoacrylate glue, compared with 3/20 (15%) who had TIPS (p ⫽ 0.005). Over the median follow-up period, the rebleed rate was 8/23 (35%) in the cyanoacrylate glue group versus 4/20 (20%) in the TIPS group (Fig. 1). Only two patients rebled as a result of shunt stenosis, and failure of stent insertion caused another rebleed. The patients who rebled after initial TIPS were managed as follows: one had re-TIPS, one required a mesocaval shunt, and the third patient responded to endoscopic therapy with cyanoacrylate glue. The fourth patient died immedi-
AJG – December, 2003
Glue vs. TIPS for Bleeding Gastric Varices
2691
Table 4. Results of X-Rays Performed After Glue Procedures Patient No.
Film
X-Ray Findings
1
CXR
2
CXR
3
AXR
4
AXR
5 6
CXR AXR
7
CXR
8 9 10
CXR CXR CXR
11
CXR
12
CXR
13
CXR
14 15 16 17
CXR CT CXR CXR
Lipiodal mass (extravascular) in fundus and in lungs Lipiodal mass (intravascular) in stomach only Lipiodal mass (intravascular) in fundus and in lungs CXR as well Large (extravascular) lipiodal mass in gastric fundus and lungs Large (extravascular) lipiodal mass Tiny lipiodol mass in fundus and lungs Lipiodol visible in fundus (extravascular) and chest Lipiodal mass in fundus and chest Lipiodal in fundus and chest Lipiodal seen in liver, varices and chest Large lipiodol seen (intravascular) and spread into abdominal vessels Lipiodal seen in fundus (intravascular) Lipiodol seen in fundus (intravascular) Trace of lipiodal seen Trace amount on CT only Lipiodal in fundus (extravascular) Trace only visible
Contrast Visible
Contrast Elsewhere
Rebleed
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y Y
N Y
N N
Y
Y
N
Y Y Y
Y Y Y
N Y N
Y
Y
Y
Y
Y
N
Y
N
N
Y Y Y Y
N N N N
N N Y N
AXR ⫽ abdominal x-ray; CT ⫽ computerized tomogram; CXR ⫽ chest x-ray; N ⫽ no; Y ⫽ yes.
ately after the rebleed, and shunt function could not be assessed. The rebleeders in the cyanoacrylate glue group had the following outcomes: four responded to repeat cyanoacrylate glue injection, two had TIPS, one had a splenectomy with gastric devascularization followed by cyanoacrylate glue injection (two episodes of rebleeding), and one died immediately. Figure 2 shows the cumulative KaplanMeier survival plot for both groups. For the period of study,
there was no significant difference in mortality rates (p ⫽ 0.95) between the two groups. Table 5 details the items that were included in the cost analysis of both procedures for a median period of 6 months. Because the total costs in both groups were not normally distributed, median costs were used for comparison. The final median cumulative cost for the follow-up period of 6 months post–index bleed was $4,138 (range, $1,618 –$25,325) for patients who received cyanoacrylate glue, compared with $11,906 (range, $6,850 –$38,110) for
Figure 1. Cumulative proportion (Kaplan-Meier plot) of patients free of rebleeding.
Figure 2. Kaplan Meier analysis of cumulative survival after the index bleed in both treatment groups.
2692
Mahadeva et al.
AJG – Vol. 98, No. 12, 2003
Table 5. Cumulative Procedures and Hospital Stay During the 6-Month Follow-up Cumulative Costing Items
TIPS (n ⫽ 20)
TIPS TIPS revision Doppler ultrasound TIPS angiogram Upper GI endoscopy Cyanoacrylate injection episodes Surgery Inpatient ward stay (day) Intensive care unit stay (day) Cumulative median cost (US$) (interquartile range)
20 2 36 4 41 1 1 357 37 $11,906 (8,200, 16,770)
patients who were treated with TIPS (p ⬍ 0.0001). The significantly higher cost of TIPS was mainly related to the cost of the procedure together with the increased length of hospitalization. This difference is clearly visible after the index bleeding episode and seems to have been sustained for up to 6 months.
DISCUSSION Cyanoacrylate glue is a safe and effective treatment for bleeding gastric varices and can be provided at a significantly lower cost when compared with TIPS placement. TIPS however, has a lower early rebleeding rate. There is little information in the current literature on the economic impact of various treatment methods for gastric varices. The higher initial cost of TIPS compared with endoscopic treatment after esophageal variceal hemorrhage is well recognized (18 –21), but TIPS has the ability to directly reduce portal hypertension and all its complications, including variceal rebleeding, with a theoretic saving in long term costs. Nevertheless, several studies that have economically evaluated TIPS against endoscopic therapy (18, 20) or surgical methods of shunting (22, 23) have failed to demonstrate a long term cost benefit. This has been attributed mostly to the high incidence of encephalopathy and shunt insufficiency in the long run post-TIPS (24), which require additional hospitalization and intervention. Our study identified readmission for complications but did not record quality-of-life measures. A recent editorial (12) suggested that endoscopic treatment with cyanoacrylate glue would probably produce equal or better results than TIPS at a lower cost, but no direct comparison studies had been performed yet to show this. Although this retrospective study has some limitations, we have demonstrated that both groups of patients with bleeding gastric varices were similar in terms of etiology of liver disease, Child-Pugh classification, and age. The disparity of initial Hb results between the two groups is of uncertain significance in the absence of any difference in hemodynamic parameters or blood product requirement. The comparison of a historical TIPS group with a more recent glue treatment group might obscure changes in practices, person-
Cyanoacrylate Glue (n ⫽ 23)
p
2 26 1 77 52 1 303 25 $4,138 (3,009, 8,290)
0.059 ns ⬍0.0001
nel, patient selection criteria, and specialty interests of the supervising clinicians. It is clear that referral rates for patients with bleeding gastric varices increased over the period of the study, from five to 12 patients per year. However, selection criteria were unchanged. The study included all patients with confirmed gastric variceal bleeding admitted during that period. Additionally, the protocol for management of these patients was unchanged during this period, including the primary agent of pharmacotherapy. Although the study numbers are small, reflecting the relatively uncommon presentation of this condition, the two groups are comparable, and valid conclusions can therefore be drawn, on the effects of the two treatment modalities. To qualify the economic evaluation further, we compared the results of both treatment modalities with recent published reports. The 6-month rebleeding rate among patients who had had TIPS for bleeding gastric varices was 26% in the study by Barange et al. (n ⫽ 32) (7) and 29% in Chau et al.’s study (n ⫽ 28) (6), compared with 20% in our group. Within a similar period, there was a comparable rate of stent stenosis in the other studies, but our 6-month mortality rate seems to be lower (43% in Chau et al. and 38% in Barange et al., respectively). Conversely, the early rebleeding rates in our patients who received cyanoacrylate glue were significantly higher than in recent reports: 35% versus 21% in Huang et al.’s study (n ⫽ 90) (13), 18.5% in the study by Lee et al. (n ⫽ 54) (14), and 15.5% in Kind et al.’s group (n ⫽ 174) (15). This might reflect a learning curve for the procedure or might be owing to the higher proportion of Child-Pugh class C patients in our cohort (compared with the Asian studies) (13, 14). A possible further limitation in this economic evaluation was the use of institutional charges as an estimate of cost of treatment. Although several other economic analyses of therapeutic trials in portal hypertension have also done the same (18, 20, 21, 23, 25), it is recognized that hospital/ institutional charges are usually different from the true costs of treatment (26). However, as this study was based in one institution, with all patients being admitted to the same unit during the period of analysis, institutional charges will be the closest estimate of economic cost. Only direct costs of both treatments were accounted for in our analysis. Because
AJG – December, 2003
the periods of hospitalization in both groups of patients were not too dissimilar, we did not expect a significant difference in indirect costs, such as time of work, etc. In conclusion, this retrospective evaluation suggests that endoscopic treatment with cyanoacrylate glue for bleeding gastric varices is less costly than TIPS. Although initial rebleed rates were higher for the glue group (7/23 vs 4/20), hemostasis was achieved after a second glue treatment in 5/7 patients in our study. Cyanoacrylate glue is therefore comparable to TIPS in terms of survival, morbidity, and complication, but clearly a prospective, randomized, controlled trial with longer term follow-up in these patients would be necessary to verify these findings. Because health economics clearly influence the clinician’s availability of therapeutic options, we agree with and support previous recommendations (12, 13) that cyanoacrylate glue should be used as first line therapy in the management of bleeding gastric varices and that TIPS should be reserved for an established failure of glue therapy.
ACKNOWLEDGMENTS We are grateful to the following for their invaluable assistance in this study: Jean Holmes, Department of Radiology, St. James’s University Hospital; the nursing staff of the Endoscopy Unit, St. James’s University Hospital; the Finance Department, Medicine, Surgery and Oncology Division, St. James’s University Hospital; and the Health Economics Research Group, Brunel University, Uxbridge, Middlesex. Reprint requests and correspondence: Charles E. Millson, St. James’s University Hospital, Liver Unit, Beckett Street, Leeds LS9, United Kingdom. Received Jan. 29, 2003; accepted July 25, 2003.
REFERENCES 1. Yassin ZM, Eita MS, Hussein A. Endoscopic sclerotherapy for bleeding gastric varices. Gut 1985;26:A1105 (abstract). 2. Trudeau W, Prindiville T. Endoscopic injection sclerosis in bleeding gastric varices. Gastrointest Endosc 1986;32:264 –8. 3. Chey WD, Elta GH. Natural history of gastric varices. Gastroenterology 1993;105:599 –602. 4. Cello J, Grendell J, Crass R, et al. Endoscopic sclerotherapy versus portacaval shunt in patients with severe cirrhosis and acute variceal haemorrhage. Long term follow-up. N Engl J Med 1987;316:11–5. 5. Orloff MJ, Bell RH, Orloff MS, et al. Prospective randomized trial of emergency portocaval shunt and emergency medical therapy in unselected cirrhotic patients with bleeding varices. Hepatology 1994;20:863–72. 6. Chau T, Patch D, Chan Y, et al. ‘Salvage’ transjugular intrahepatic portosystemic shunts: Gastric fundal compared with oesophageal variceal bleeding. Gastroenterology 1998;114: 981–7.
Glue vs. TIPS for Bleeding Gastric Varices
2693
7. Barange K, Peron JM, Imani K, et al. Transjugular intrahepatic portosystemic shunt in the treatment of refractory bleeding from ruptured gastric varices. Hepatology 1990;30:1139 –43. 8. Helton WS, Belshaw A, Althaus S, et al. Critical appraisal of the angiographic portacaval shunt (TIPS). Am J Surg 1993; 165:566 –71. 9. Chalasani N, Clark WS, Martin LG, et al. Determinants of mortality in patients with advanced cirrhosis after tranjugular intrahepatic portosystemic shunting. Gastroenterology 2000; 118:138 –44. 10. Spahr L, Dufresne MP, Bui BT. Efficacy of TIPS in the prevention of re-bleeding from oesophageal and fundal varices: A comparative study. Hepatology 1995;22:296A (abstract). 11. Sarin SK, Primignani M, Agarwal SR. 3rd Baveno International Consensus Workshop on Portal Hypertension. Oxford: Blackwell Science, 2001:76 –96. 12. Binmoeller KF. Glue for gastric varices: Some sticky issues. Gastrointest Endosc 2000;52:298⫺301. 13. Huang YH, Yeh HZ, Chen GH, et al. Endoscopic treatment of bleeding gastric varices by N-butyl-2-cyanoacrylate (Histoacryl) injection: Long-term efficacy and safety. Gastrointest Endosc 2000;52:160 –7. 14. Lee YT, Chan FKL, Ng EKW, et al. EUS-guided injection of cyanoacrylate for bleeding gastric varices. Gastrointest Endosc 2000;52:168 –74. 15. Kind R, Guglielmi A, Rodella L, et al. Bucrylate treatment of bleeding gastric varices: 12 years’ experience. Endoscopy 2000;32:512–9. 16. Rossle M, Haag K, Ochs A, et al. Transjugular intrahepatic portosystemic stent-shunt procedure for variceal bleeding. N Engl J Med 1994;330:165–71. 17. Sarin SK, Kumar A. Gastric varices: Profile, classification and management. Am J Gastroenterol 1989;84:1244 –9. 18. Cello JP, Ring EJ, Olkott EW, et al. TIPS vs sclerotherapy for variceal haemorrhage. Ann Intern Med 1997;126:858 –67. 19. Jalan R, Forrest EH, Stanley AJ, et al. A randomised trial comparing TIPS with variceal band ligation in the prevention of re-bleeding from oesophageal varices. Hepatology 1997; 26:1115–22. 20. Meddi P, Merli M, Lionetti R, et al. Cost analysis for the prevention of variceal bleeding: A comparison between transjugular intrahepatic portosystemic shunt and endoscopic sclerotherapy in a selected group of Italian cirrhotic patients. Hepatology 1999;29:1074 –7. 21. Russo MW, Zacks SL, Sandler RS, Brown RSJ. Cost-effectiveness analysis of transjugular intrahepatic portosystemic shunt (TIPS) versus endoscopic therapy for the prevention of recurrent esophageal variceal bleeding. Hepatology 2000;31: 358 –63. 22. Rosemurgy AS, Bloomston M, Zervos EE, et al. TIPS versus H-graft portacaval shunt in the management of bleeding varices: a cost-benefit analysis. Surgery 1997;122:794 –9. 23. Zacks SL, Sandler RS, Biddle AK, et al. Decision-analysis of transjugular intrahepatic portosystemic shunt versus distal splenorenal shunt for portal hypertension. Hepatology 1999; 29:1399⫺1405. 24. Grace ND. TIPS: The long and the short of it. Gastroenterology 1997;112:1040 –3. 25. Chung R, Lewis JW. Cost of treatment of bleeding esophageal varices. Arch Surg 1983;118:482–5. 26. Finkler SA. The distinction between cost and charges. Ann Intern Med 1982;96:102–9.
Vol. 98, No. 12, 2003 ISSN 0002-9270/03/$30.00 doi:10.1016/j.amjgastroenterol.2003.09.039
Cost-Effectiveness of N-Butyl-2-Cyanoacrylate (Histoacryl) Glue Injections Versus Transjugular Intrahepatic Portosystemic Shunt in the Management of Acute Gastric Variceal Bleeding S. Mahadeva, M.B., M.R.C.P., M. C. Bellamy, M.A., M.B., F.R.C.A., D. Kessel, M.A., M.R.C.P., F.R.C.R., M. H. Davies, M.D., F.R.C.P., and C. E. Millson, M.D., F.R.C.P. Departments of Hepatology and Liver Transplantation, Intensive Care Medicine, and Radiology, St. James’s University Hospital, Leeds, United Kingdom
ABSTRACT The management of bleeding gastric varices has not been standardized. Although transjugular intrahepatic portosystemic shunt (TIPS) is used in most centers, endoscopic treatment with N-butyl-2-cyanoacrylate (cyanoacrylate) glue has recently been shown to be effective. Cost-effectiveness analyses of these methods are lacking. METHODS: We performed a retrospective review of patients with bleeding gastric varices treated either by TIPS or cyanoacrylate glue injection. Economic analysis was based on direct costs for a fixed financial year. The two groups were compared for a period of 6 months follow-up, to liver transplantation, or death for each patient. RESULTS: Between January, 1995 and December, 1999, 20 patients with bleeding gastric varices had TIPS; 23 patients had cyanoacrylate glue injection from January, 2000 to October, 2001. There were no significant differences between the two groups in patient characteristics, transfusion requirement, and gastric variceal anatomy. In the TIPS group, 15/20 patients had the procedure performed within 24 h of hemorrhage, and 90% of stent insertions were successful. Complications consisted of two cases of pulmonary edema, two cases of severe encephalopathy, and a 15% stenosis rate at 6 months. In the glue group, there were 3 ⫾ 1.5 endoscopies and 2 ⫾ 1 injections per patient, with a 96% initial hemostasis. There was one case of (glue) pulmonary embolism and one blocked front endoscope lens, which required repair. The initial rebleed rate was significantly lower in patients who had TIPS (15% vs 30%, p ⫽ 0.005). The inpatient stay was shorter in the glue group (13 ⫾ 1 vs 18 ⫾ 2 days, p ⫽ 0.05), but there was no difference in the overall mortality rate. The median cost within 6 months of initial gastric variceal bleeding was $4,138 ($3,009 –$8,290) for glue versus $11,906 ($8,200 –$16,770) for TIPS (p ⬍ 0.0001). CONCLUSION: In this comparable group of patients, cyanoacrylate glue injection was more cost effective than TIPS in
the management of acute gastric variceal bleeding. A prospective, randomized trial would be required to confirm our analysis. (Am J Gastroenterol 2003;98:2688⫺2693. © 2003 by Am. Coll. of Gastroenterology)
INTRODUCTION The optimal management of bleeding gastric varices remains uncertain. Standard endoscopic techniques do not seem to be effective (1–3), and surgical shunting is associated with a high mortality (4, 5); hence, transjugular intrahepatic portosystemic shunt (TIPS) has emerged as the recommended treatment. Although TIPS achieves hemostasis in more than 90% of bleeding gastric varices (6, 7), the early mortality rate for emergently placed TIPS in patients with advanced liver disease has been reported to be between 40% and 60% (8, 9), and recurrent bleeding ranges from 29% (6) to 53% (10). Recently, endoscopic treatment with cyanoacrylate glue has been reported as an effective option in the management of bleeding gastric varices (11, 12). Large series from Asia and Europe have shown primary hemostasis rates of 94 – 97%, low rebleeding rates, and early mortality rates of only 12–20% (13–15), prompting several authors to propose cyanoacrylate glue as the first choice for bleeding gastric varices (11–13). Because neither TIPS nor endoscopic cyanoacrylate injection actually treats the underlying cause of the gastric varices, the effectiveness and economic impact of the technique should influence the clinician’s choice of therapy. It has been suggested that endoscopic treatment with cyanoacrylate is probably cheaper than TIPS (12), but this has not been proved to date. The aim of this study was to compare, retrospectively, the outcomes and health care costs of patients treated with either cyanoacrylate glue or TIPS as the primary treatment. Costs were evaluated with an intent-to-treat analysis and included rescue therapies for both modalities in the evaluation. Out-
AJG – December, 2003
come measures included rebleed rates, transfusion requirement, and survival.
Glue vs. TIPS for Bleeding Gastric Varices
Table 1. Institutional Charges as a Measure of Direct Costs of the Parameters Analyzed Cost Items
MATERIALS AND METHODS All patients referred to this institution for bleeding gastric varices from 1995 were identified from our TIPS database (department of vascular radiology) and endoscopy unit (Endoscribe database, Astra, Endoscribe, Medlboss, Adelaide, Australia). Only patients with confirmed bleeding gastric varices on upper GI endoscopy were included in the study. This was defined as evidence of visible or active bleeding at the time of endoscopy, the presence of fresh clot on varices, or the presence of “red spots” on gastric varices with fresh or altered blood in the stomach without any other possible source of bleeding. Relevant clinical data were reviewed from patients’ clinical records, and endoscopic data were recorded from clinical notes and computerized endoscopy reports. TIPS had been performed under general anesthesia with a standard technique described previously (16). Only Wallstents (Pfeizer, Bulach, Switzerland) or Memotherm (Angiomed, Karlsruhe, Germany) stents had been used. After insertion, routine Doppler ultrasound scanning would be performed after 2 days and after 1 wk and then on an every-3-months basis to assess stent patency. If shunt dysfunction was suspected on Doppler scan, angiography would be performed. The technique of endoscopic cyanoacrylate glue injection used in our unit has been described previously (13). At endoscopy, N-butyl-2-cyanoacrylate (Histoacryl, B. Braun, Melsungen, Germany) was diluted with Lipiodol (Laboratoire Guerbet, Aulnay-Sousels, France) in a 1:1.5 ratio and injected as a bolus of 1 to 2 ml, according to variceal size. Most patients had a plain abdominal x-ray postendoscopy to evaluate opacification of the varices. Follow-up post–index endoscopy was arranged within 48 h, then on a weekly or monthly basis, depending on the degree of variceal obliteration. The latter was assessed by inspection and blunt prodding of the variceal mass at endoscopy. Because the patients who had been treated with cyanoacrylate had a shorter follow-up period (median of 6 months) than those originally treated with TIPS, cost comparison between the two groups of patients were limited to 6 months post–index bleed, time of liver transplantation, or death. During the period studied, days of hospitalization owing to the initial index bleeding episode and further hospitalizations for rebleeding or complications of treatment were recorded as the cost of the therapeutic procedures. The cost of TIPS included all equipment, time of medical and radiologic staff, medication, and allotted 2 h for general anesthesia. Similarly, the cost of endoscopic cyanoacrylate injection involved all equipment, the time of medical and nursing staff, and use of the endoscopy unit sessions. Estimates of inpatient stay unit costs include nursing staff costs, administrative and clerical staff costs, consumables, equipment, overheads, and capital costs. We assumed that
2689
Inpatient stay (per/day) Liver ward General medical ward Intensive care unit Radiology TIPS TIPS revision TIPS venogram Ultrasound scan ⫹ Doppler Endoscopy Upper GI endoscopy Upper GI endoscopy ⫹ glue (5 injections) Surgery Mesocaval shunt Splenectomy and gastric devascularisation
US$ 195.2 101.9 2159 5080 2030 610.16 102.45 150.80 389.38 8027 6291
there would be no differences in ward staff fee, routine blood investigations, standard vasoactive drugs, and basic radiology (i.e., chest x-ray) between the two groups because they had been managed in the same hospital with little difference in practice apart from the principle therapeutic modality. All costs were based on the Healthcare Resource Grouping codes (National Health Service, United Kingdom) of this institution over the April, 2000 to March, 2001 financial period (Table 1). Cost analysis of each therapeutic modality was calculated on an intent-to-treat basis. Survival and rebleeding were calculated according to the Kaplan-Meier method, and differences between treatment groups were analyzed by the log-rank test. Quantitative and qualitative variables were compared with either MannWhitney U test or Fisher exact test, where appropriate. The cost difference for the period of study was analyzed with the Mann-Whitney U test. Statistical significance was assumed at p ⬍ 0.05.
RESULTS Between January, 1995 and December, 1999, 20 patients underwent TIPS insertion for confirmed bleeding gastric varices. Between January, 2000 and October, 2001, a further 23 patients with bleeding gastric varices were treated with cyanoacrylate glue. There were no significant differences between the two treatment groups in terms of patient demographics, etiology of liver disease, Child-Pugh classification, and gastric variceal anatomy (Table 2). The latter had been based on the classification system used by Sarin et al. (17). To assess the severity of bleeding in both groups, hemodynamic and hematological parameters were compared (Table 3). There was no significant hemodynamic difference between the two groups of patients on the day of admission during the index bleed, but there was a higher mean serum Hb (8.7 ⫾ 0.46 vs 7.6 ⫾ 0.35 g/dl, p ⫽ 0.038) and a slightly lower transfusion requirement (8 ⫾ 1 vs 9 ⫾ 2 U; mean difference 2.3, 95% CI ⫽ ⫺1.8 to 6.4) in the
2690
Mahadeva et al.
AJG – Vol. 98, No. 12, 2003
Table 2. Characteristics of Patients in the Study
Age (yr) Sex (male/female) Etiology Alcohol Cholestatic Cryptogenic Others Child-Pugh class A B C GV anatomy GOV 1 GOV 2
TIPS (n ⫽ 20)
Cyanoacrylate Glue (n ⫽ 23)
p
52 ⫾ 3 13/7
55 ⫾ 3 15/8
ns* ns†
12 3 2 3
14 3 3 3
ns†
4 6 10
4 8 9
ns†
13 7
15 8
ns†
GOV ⫽ gastroesophageal varices; GV ⫽ gastric varices. * Mann-Whitney U test. † Fisher exact test.
group treated with cyanoacrylate. At the index endoscopy, active bleeding was present in nine patients who received cyanoacrylate glue, compared with 12 patients in the TIPS group; a fresh adherent clot was seen in six patients in the glue group, compared with eight patients who had TIPS (p ⫽ ns). In the TIPS group, 15/20 patients had the procedure within 24 h of hemorrhage. Endoscopic treatment (two with thrombin, two with sodium tetradecyl sulfate) was unsuccessfully attempted in a further four patients before stent insertion. Eleven patients required either a SengstakenBlakemore or Minnesota tube for stabilization as well. Out of 20 stent insertions, 18 (90%) were successful. In the remaining two, shunting was unsuccessful despite two stent deployments in each of both cases; one patient rebled and eventually required a mesocaval shunt. The mean inferior Table 3. Hematological and Hemodynamic Parameters in Patients (Mean ⫾ SEM)
Hemoglobin (g/dL)† Heart rate† Systolic blood pressure (mm Hg)† Blood transfused (U) FFP transfused (U)
TIPS (n ⫽ 20)
Cyanoacrylate Glue (n ⫽ 23)
p*
7.6 ⫾ 0.35
8.7 ⫾ 0.46
0.038
90 ⫾ 4.2 113 ⫾ 7
94 ⫾ 3.4 117 ⫾ 4
ns ns
9⫾2
8⫾1
ns
6⫾1
5⫾1
ns
FFP ⫽ fresh frozen plasma. * By Mann-Whitney U test. † Measurements made on admission to ward.
vena cava–portal vein gradient dropped from 21 ⫾ 2 mm Hg to 5 ⫾ 0.7 mm Hg after TIPS insertion (p ⬍ 0.001). One patient continued to bleed despite shunting, and embolization of the varices was performed at the same time as TIPS. The median follow-up period among all cases was 12 months (range 0.1– 60). Early complications occurred in four cases: two with severe encephalopathy and two with pulmonary edema. All but one responded to medical therapy. The TIPS stenosis rate was 15% (three patients) over the 6-month cost analysis period; two presented with recurrent variceal bleeding, and the third was detected during routine surveillance. Further stent deployment was performed in two of these patients. Two patients had a liver transplant 8 days and 6 wk after the index bleed, respectively. Both had survived up to the follow-up period. In the group of patients that received cyanoacrylate glue, 10 patients required a Minnesota tube for stabilization. Out of 23 patients, 20 received cyanoacrylate glue injection within 24 h. Primary hemostasis was achieved in 96% of cases. The mean number of endoscopies with cyanoacrylate glue injection, to achieve variceal obliteration, was 2 ⫾ 1 (range 1–5) per patient. A mean of 7.5 ml (three vials) of cyanoacrylate was used during each endoscopic gluing session. Over the 6-month period, there were a mean of 3 ⫾ 1.5 (range 2–7) endoscopies performed per patient. The median follow-up period was 6 months (range 0.25–20 months). Post– glue injection abdominal or chest x-rays were performed in 17 patients (Table 4). Of these, seven had glue distribution that clearly appeared intravascular, and six appeared extravascular or intramural. Two had contrast in an intramural distribution but had varices outlined as well. There was no correlation between distribution of glue on x-ray and rebleed tendency. Ten had evidence of glue in the chest as well as the injection site in the gastric fundus. Whether this glue was transmitted by a vascular route or by aspiration of remnant glue/mucus mix from the stomach is uncertain. One patient with evidence of glue within the lung died of respiratory failure 24 h after the procedure. However, aspiration pneumonia is common in this group of patients, and glue was clearly visible in a significant percentage of patients receiving cyanoacrylate glue. A further complication consisted of a single case of occlusion of an endoscope (Olympus GIF IT240, Olympus, Southend-onSea, UK) front lens, which required repairing by the manufacturers. The cost of this was included in our analysis. Early (defined as less than 30 days) rebleeding occurred in 7/23 patients (30%) who had cyanoacrylate glue, compared with 3/20 (15%) who had TIPS (p ⫽ 0.005). Over the median follow-up period, the rebleed rate was 8/23 (35%) in the cyanoacrylate glue group versus 4/20 (20%) in the TIPS group (Fig. 1). Only two patients rebled as a result of shunt stenosis, and failure of stent insertion caused another rebleed. The patients who rebled after initial TIPS were managed as follows: one had re-TIPS, one required a mesocaval shunt, and the third patient responded to endoscopic therapy with cyanoacrylate glue. The fourth patient died immedi-
AJG – December, 2003
Glue vs. TIPS for Bleeding Gastric Varices
2691
Table 4. Results of X-Rays Performed After Glue Procedures Patient No.
Film
X-Ray Findings
1
CXR
2
CXR
3
AXR
4
AXR
5 6
CXR AXR
7
CXR
8 9 10
CXR CXR CXR
11
CXR
12
CXR
13
CXR
14 15 16 17
CXR CT CXR CXR
Lipiodal mass (extravascular) in fundus and in lungs Lipiodal mass (intravascular) in stomach only Lipiodal mass (intravascular) in fundus and in lungs CXR as well Large (extravascular) lipiodal mass in gastric fundus and lungs Large (extravascular) lipiodal mass Tiny lipiodol mass in fundus and lungs Lipiodol visible in fundus (extravascular) and chest Lipiodal mass in fundus and chest Lipiodal in fundus and chest Lipiodal seen in liver, varices and chest Large lipiodol seen (intravascular) and spread into abdominal vessels Lipiodal seen in fundus (intravascular) Lipiodol seen in fundus (intravascular) Trace of lipiodal seen Trace amount on CT only Lipiodal in fundus (extravascular) Trace only visible
Contrast Visible
Contrast Elsewhere
Rebleed
Y
Y
Y
Y
N
Y
Y
Y
Y
Y
Y
Y
Y Y
N Y
N N
Y
Y
N
Y Y Y
Y Y Y
N Y N
Y
Y
Y
Y
Y
N
Y
N
N
Y Y Y Y
N N N N
N N Y N
AXR ⫽ abdominal x-ray; CT ⫽ computerized tomogram; CXR ⫽ chest x-ray; N ⫽ no; Y ⫽ yes.
ately after the rebleed, and shunt function could not be assessed. The rebleeders in the cyanoacrylate glue group had the following outcomes: four responded to repeat cyanoacrylate glue injection, two had TIPS, one had a splenectomy with gastric devascularization followed by cyanoacrylate glue injection (two episodes of rebleeding), and one died immediately. Figure 2 shows the cumulative KaplanMeier survival plot for both groups. For the period of study,
there was no significant difference in mortality rates (p ⫽ 0.95) between the two groups. Table 5 details the items that were included in the cost analysis of both procedures for a median period of 6 months. Because the total costs in both groups were not normally distributed, median costs were used for comparison. The final median cumulative cost for the follow-up period of 6 months post–index bleed was $4,138 (range, $1,618 –$25,325) for patients who received cyanoacrylate glue, compared with $11,906 (range, $6,850 –$38,110) for
Figure 1. Cumulative proportion (Kaplan-Meier plot) of patients free of rebleeding.
Figure 2. Kaplan Meier analysis of cumulative survival after the index bleed in both treatment groups.
2692
Mahadeva et al.
AJG – Vol. 98, No. 12, 2003
Table 5. Cumulative Procedures and Hospital Stay During the 6-Month Follow-up Cumulative Costing Items
TIPS (n ⫽ 20)
TIPS TIPS revision Doppler ultrasound TIPS angiogram Upper GI endoscopy Cyanoacrylate injection episodes Surgery Inpatient ward stay (day) Intensive care unit stay (day) Cumulative median cost (US$) (interquartile range)
20 2 36 4 41 1 1 357 37 $11,906 (8,200, 16,770)
patients who were treated with TIPS (p ⬍ 0.0001). The significantly higher cost of TIPS was mainly related to the cost of the procedure together with the increased length of hospitalization. This difference is clearly visible after the index bleeding episode and seems to have been sustained for up to 6 months.
DISCUSSION Cyanoacrylate glue is a safe and effective treatment for bleeding gastric varices and can be provided at a significantly lower cost when compared with TIPS placement. TIPS however, has a lower early rebleeding rate. There is little information in the current literature on the economic impact of various treatment methods for gastric varices. The higher initial cost of TIPS compared with endoscopic treatment after esophageal variceal hemorrhage is well recognized (18 –21), but TIPS has the ability to directly reduce portal hypertension and all its complications, including variceal rebleeding, with a theoretic saving in long term costs. Nevertheless, several studies that have economically evaluated TIPS against endoscopic therapy (18, 20) or surgical methods of shunting (22, 23) have failed to demonstrate a long term cost benefit. This has been attributed mostly to the high incidence of encephalopathy and shunt insufficiency in the long run post-TIPS (24), which require additional hospitalization and intervention. Our study identified readmission for complications but did not record quality-of-life measures. A recent editorial (12) suggested that endoscopic treatment with cyanoacrylate glue would probably produce equal or better results than TIPS at a lower cost, but no direct comparison studies had been performed yet to show this. Although this retrospective study has some limitations, we have demonstrated that both groups of patients with bleeding gastric varices were similar in terms of etiology of liver disease, Child-Pugh classification, and age. The disparity of initial Hb results between the two groups is of uncertain significance in the absence of any difference in hemodynamic parameters or blood product requirement. The comparison of a historical TIPS group with a more recent glue treatment group might obscure changes in practices, person-
Cyanoacrylate Glue (n ⫽ 23)
p
2 26 1 77 52 1 303 25 $4,138 (3,009, 8,290)
0.059 ns ⬍0.0001
nel, patient selection criteria, and specialty interests of the supervising clinicians. It is clear that referral rates for patients with bleeding gastric varices increased over the period of the study, from five to 12 patients per year. However, selection criteria were unchanged. The study included all patients with confirmed gastric variceal bleeding admitted during that period. Additionally, the protocol for management of these patients was unchanged during this period, including the primary agent of pharmacotherapy. Although the study numbers are small, reflecting the relatively uncommon presentation of this condition, the two groups are comparable, and valid conclusions can therefore be drawn, on the effects of the two treatment modalities. To qualify the economic evaluation further, we compared the results of both treatment modalities with recent published reports. The 6-month rebleeding rate among patients who had had TIPS for bleeding gastric varices was 26% in the study by Barange et al. (n ⫽ 32) (7) and 29% in Chau et al.’s study (n ⫽ 28) (6), compared with 20% in our group. Within a similar period, there was a comparable rate of stent stenosis in the other studies, but our 6-month mortality rate seems to be lower (43% in Chau et al. and 38% in Barange et al., respectively). Conversely, the early rebleeding rates in our patients who received cyanoacrylate glue were significantly higher than in recent reports: 35% versus 21% in Huang et al.’s study (n ⫽ 90) (13), 18.5% in the study by Lee et al. (n ⫽ 54) (14), and 15.5% in Kind et al.’s group (n ⫽ 174) (15). This might reflect a learning curve for the procedure or might be owing to the higher proportion of Child-Pugh class C patients in our cohort (compared with the Asian studies) (13, 14). A possible further limitation in this economic evaluation was the use of institutional charges as an estimate of cost of treatment. Although several other economic analyses of therapeutic trials in portal hypertension have also done the same (18, 20, 21, 23, 25), it is recognized that hospital/ institutional charges are usually different from the true costs of treatment (26). However, as this study was based in one institution, with all patients being admitted to the same unit during the period of analysis, institutional charges will be the closest estimate of economic cost. Only direct costs of both treatments were accounted for in our analysis. Because
AJG – December, 2003
the periods of hospitalization in both groups of patients were not too dissimilar, we did not expect a significant difference in indirect costs, such as time of work, etc. In conclusion, this retrospective evaluation suggests that endoscopic treatment with cyanoacrylate glue for bleeding gastric varices is less costly than TIPS. Although initial rebleed rates were higher for the glue group (7/23 vs 4/20), hemostasis was achieved after a second glue treatment in 5/7 patients in our study. Cyanoacrylate glue is therefore comparable to TIPS in terms of survival, morbidity, and complication, but clearly a prospective, randomized, controlled trial with longer term follow-up in these patients would be necessary to verify these findings. Because health economics clearly influence the clinician’s availability of therapeutic options, we agree with and support previous recommendations (12, 13) that cyanoacrylate glue should be used as first line therapy in the management of bleeding gastric varices and that TIPS should be reserved for an established failure of glue therapy.
ACKNOWLEDGMENTS We are grateful to the following for their invaluable assistance in this study: Jean Holmes, Department of Radiology, St. James’s University Hospital; the nursing staff of the Endoscopy Unit, St. James’s University Hospital; the Finance Department, Medicine, Surgery and Oncology Division, St. James’s University Hospital; and the Health Economics Research Group, Brunel University, Uxbridge, Middlesex. Reprint requests and correspondence: Charles E. Millson, St. James’s University Hospital, Liver Unit, Beckett Street, Leeds LS9, United Kingdom. Received Jan. 29, 2003; accepted July 25, 2003.
REFERENCES 1. Yassin ZM, Eita MS, Hussein A. Endoscopic sclerotherapy for bleeding gastric varices. Gut 1985;26:A1105 (abstract). 2. Trudeau W, Prindiville T. Endoscopic injection sclerosis in bleeding gastric varices. Gastrointest Endosc 1986;32:264 –8. 3. Chey WD, Elta GH. Natural history of gastric varices. Gastroenterology 1993;105:599 –602. 4. Cello J, Grendell J, Crass R, et al. Endoscopic sclerotherapy versus portacaval shunt in patients with severe cirrhosis and acute variceal haemorrhage. Long term follow-up. N Engl J Med 1987;316:11–5. 5. Orloff MJ, Bell RH, Orloff MS, et al. Prospective randomized trial of emergency portocaval shunt and emergency medical therapy in unselected cirrhotic patients with bleeding varices. Hepatology 1994;20:863–72. 6. Chau T, Patch D, Chan Y, et al. ‘Salvage’ transjugular intrahepatic portosystemic shunts: Gastric fundal compared with oesophageal variceal bleeding. Gastroenterology 1998;114: 981–7.
Glue vs. TIPS for Bleeding Gastric Varices
2693
7. Barange K, Peron JM, Imani K, et al. Transjugular intrahepatic portosystemic shunt in the treatment of refractory bleeding from ruptured gastric varices. Hepatology 1990;30:1139 –43. 8. Helton WS, Belshaw A, Althaus S, et al. Critical appraisal of the angiographic portacaval shunt (TIPS). Am J Surg 1993; 165:566 –71. 9. Chalasani N, Clark WS, Martin LG, et al. Determinants of mortality in patients with advanced cirrhosis after tranjugular intrahepatic portosystemic shunting. Gastroenterology 2000; 118:138 –44. 10. Spahr L, Dufresne MP, Bui BT. Efficacy of TIPS in the prevention of re-bleeding from oesophageal and fundal varices: A comparative study. Hepatology 1995;22:296A (abstract). 11. Sarin SK, Primignani M, Agarwal SR. 3rd Baveno International Consensus Workshop on Portal Hypertension. Oxford: Blackwell Science, 2001:76 –96. 12. Binmoeller KF. Glue for gastric varices: Some sticky issues. Gastrointest Endosc 2000;52:298⫺301. 13. Huang YH, Yeh HZ, Chen GH, et al. Endoscopic treatment of bleeding gastric varices by N-butyl-2-cyanoacrylate (Histoacryl) injection: Long-term efficacy and safety. Gastrointest Endosc 2000;52:160 –7. 14. Lee YT, Chan FKL, Ng EKW, et al. EUS-guided injection of cyanoacrylate for bleeding gastric varices. Gastrointest Endosc 2000;52:168 –74. 15. Kind R, Guglielmi A, Rodella L, et al. Bucrylate treatment of bleeding gastric varices: 12 years’ experience. Endoscopy 2000;32:512–9. 16. Rossle M, Haag K, Ochs A, et al. Transjugular intrahepatic portosystemic stent-shunt procedure for variceal bleeding. N Engl J Med 1994;330:165–71. 17. Sarin SK, Kumar A. Gastric varices: Profile, classification and management. Am J Gastroenterol 1989;84:1244 –9. 18. Cello JP, Ring EJ, Olkott EW, et al. TIPS vs sclerotherapy for variceal haemorrhage. Ann Intern Med 1997;126:858 –67. 19. Jalan R, Forrest EH, Stanley AJ, et al. A randomised trial comparing TIPS with variceal band ligation in the prevention of re-bleeding from oesophageal varices. Hepatology 1997; 26:1115–22. 20. Meddi P, Merli M, Lionetti R, et al. Cost analysis for the prevention of variceal bleeding: A comparison between transjugular intrahepatic portosystemic shunt and endoscopic sclerotherapy in a selected group of Italian cirrhotic patients. Hepatology 1999;29:1074 –7. 21. Russo MW, Zacks SL, Sandler RS, Brown RSJ. Cost-effectiveness analysis of transjugular intrahepatic portosystemic shunt (TIPS) versus endoscopic therapy for the prevention of recurrent esophageal variceal bleeding. Hepatology 2000;31: 358 –63. 22. Rosemurgy AS, Bloomston M, Zervos EE, et al. TIPS versus H-graft portacaval shunt in the management of bleeding varices: a cost-benefit analysis. Surgery 1997;122:794 –9. 23. Zacks SL, Sandler RS, Biddle AK, et al. Decision-analysis of transjugular intrahepatic portosystemic shunt versus distal splenorenal shunt for portal hypertension. Hepatology 1999; 29:1399⫺1405. 24. Grace ND. TIPS: The long and the short of it. Gastroenterology 1997;112:1040 –3. 25. Chung R, Lewis JW. Cost of treatment of bleeding esophageal varices. Arch Surg 1983;118:482–5. 26. Finkler SA. The distinction between cost and charges. Ann Intern Med 1982;96:102–9.