Marco Ballarè

Hepatic clearance of D-sorbitol: noninvasive test for evaluating functional liver plasma flow

The hepatic clearance of D-sorbitol, a natural polyol which is metabolized by the liver, was studied in normal and cirrhotic subjects after bolus intravenous injection (2 g) and during constant infusion (54 mg/min) with the aim of providing a noninvasive and simple measure of functional liver plasma flow. The high hepatic extraction of D-sorbitol and the dose-independence of its clearance pointed to a flow-dependent clearance regimen. The renal excretion was taken into account when computing the hepatic clearance. Day-to-day reproducibility of the test was good. No significant difference was found when the hepatic clearance was measured by bolus injection or constant infusion methods. As measured by the bolus injection method, the mean (+/- SD) hepatic clearance in the normal subjects (911 +/- 137 ml/min) was significantly greater (P less than 0.001) than that of the cirrhotics (456 +/- 181 ml/min).The hepatic clearance ofd-sorbitol, a natural polyol which is metabolized by the liver, was studied in normal and cirrhotic subjects after bolus intravenous injection (2 g) and during constant infusion (54 mg/min) with the aim of providing a noninvasive and simple measure of functional liver plasma flow. The high hepatic extraction ofd-sorbitol and the dose-independence of its clearance pointed to a flow-dependent clearance regimen. The renal excretion was taken into account when computing the hepatic clearance. Day-today reproducibility of the test was good. No significant difference was found when the hepatic clearance was measured by bolus injection or constant infusion methods. As measured by the bolus injection method, the mean (±sd) hepatic clearance in the normal subjects (911±137 ml/min) was significantly greater (P<0.001) than that of the cirrhotics (456±181 ml/min).

Digestive Endoscopy Is Not a Major Risk Factor for Transmitting Hepatitis C Virus

Context Controversy persists regarding the risk for transmission of hepatitis C virus (HCV) as a result of digestive-tract endoscopy. Contribution This prospective study of HCV-negative patients who underwent gastroscopy with the same endoscopes as HCV-positive patients showed no transmission of infection on follow-up 6 months later. Biopsy with reusable or disposable forceps did not increase the risk for HCV infection. Blood donors who were HCV negative without endoscopic exposure showed a few conversions to infected status an average of 2.5 years later. Implications The risk for HCV transmission by endoscopy is extremely low when standard instrument-cleaning techniques are used. The Editors Health carerelated procedures have been implicated in the transmission of a consistent proportion of contemporary hepatitis C virus (HCV) infections. The role of major surgical operations, such as cardiovascular, gynecologic, and orthopedic procedures, is well established. However, the role of less invasive procedures, such as digestive endoscopy, remains a matter of debate. A claim from a retrospective French study (1) that digestive endoscopic procedures are a major cause of HCV transmission among blood donors has not been substantiated by other authors (2, 3); acquisition of HCV through endoscopy has in fact been rarely reported in recent years (4, 5). Nevertheless, endoscopy as a vehicle for HCV transmission has been suspected since 1996, when blood banks in France and Italy suspended donors who reported a history of recent digestive endoscopy from donating blood for 6 months and up to 1 year, respectively. It is therefore important to establish whether digestive endoscopy represents a real risk and, if so, to define its magnitude. We conducted a prospective study among outpatients referred to 3 endoscopic units in northwestern Italy from 1999 to 2002. The patients entering the study were tested for antibody to HCV (anti-HCV) at baseline and 6 months after endoscopy. The incidence of HCV infection in this cohort was compared with that in blood donors recruited in the same area and during the same time period; these donors had not undergone any digestive endoscopic procedure. Methods Endoscopy Cohort Between January 1999 and December 2002, all of the outpatients referred for upper digestive endoscopy to 3 endoscopic units in northwestern Italy (1 secondary referral center and 2 tertiary referral centers) were asked to participate in this study. Eligibility criteria were age older than 18 years and indication for gastroscopy. We restricted the procedure to gastroscopy in order to obtain a high rate of invasive procedures (for example, gastric biopsy). We excluded patients if they were hospitalized, had previously undergone endoscopic procedures, were known anti-HCV carriers, or had to undergo additional endoscopic procedures other than gastroscopy. However, to identify the potentially infective population, we retrospectively looked for known HCV carriers who underwent gastroscopy in the 3 centers between January 1999 and December 2002. Of 11348 patients fulfilling the inclusion criteria, 9188 (81.0%) agreed to participate and gave written consent. They completed a questionnaire about risk factors for HCV infection during the past 6 months, and a serum sample was obtained from each immediately before endoscopy. Mild sedation with midazolam and hyoscine butylbromide was administered to each patient by using disposable syringes and vials. Gastroscopies were done by using various types of endoscopes, including fiberscopes and video endoscopes (Olympus GIF-Q20, GIF-Q30, GIF-IT30, GIF-IT140, Olympus Europe, Hamburg, Germany). Biopsies were performed with disposable biopsy forceps (Radial Jaw 3, Boston Scientific Microvasive, Natick, Massachusetts) in one center and reusable biopsy forceps (FB-24U-1, Olympus Europe) in another center; the third center used reusable forceps (EN-62143, Pescetto, Genova, Italy) in 1999 and disposable forceps (Max Capacity, Boston Scientific Microvasive) after 1999. Each patient was invited to attend a follow-up visit 6 months after endoscopy in order to obtain a serum sample for determining anti-HCV; at this visit, the patient was asked to complete the HCV questionnaire again. To reduce the risk for false-negative results, potentially immunodeficient patients (those undergoing hemodialysis or receiving immunosuppressive treatment) were also tested for HCV RNA by polymerase chain reaction (PCR). All patients who did not attend the follow-up visit were recontacted by telephone. Among patients in the endoscopy cohort, we identified an at-risk subset of patients: Overall, 912 endoscopic procedures (732 gastroscopies performed on known HCV carriers and 180 gastroscopies performed on newly discovered HCV carriers) were considered potentially infective. When we considered that each endoscope was used 3 times during the endoscopic session and assumed that the anti-HCVpositive patient was the first, second, or third at random, the number of exposed patients per HCV-infectedpatient-day was 0, 1, or 2, with equal probability (the mean of those numbers is 1). Blood Donors Cohort Using a computerized database, we retrospectively identified all 51645 consecutive blood donors at 2 transfusion centers in Torino and Pinerolo between January 1999 and December 2002 who were negative for HCV. Of these, 415 (0.8%) reported previous digestive endoscopy; the blood bank database did not record invasive procedures performed during endoscopy (such as biopsy and polypectomy). These 415 donors were asked to repeat the serologic and virologic HCV tests: 329 (79.3%) agreed, and 86 declined. Of the 51230 blood donors who did not undergo endoscopic procedures during the observation period, 38 280 (74.7%) were tested again after a mean of 2.49 years (range, 6 months to 4 years); the remaining 12 950 blood donors could not be contacted by telephone for retesting or declined to be retested. Retested blood donors found to be newly positive for anti-HCV completed a structured questionnaire aimed at investigating risk factors for HCV infection, including endoscopic procedures, travel history, sexual activity, and potential parenteral exposures to blood or blood products (previous blood, platelet, or plasma transfusions; administration of coagulation factor concentrates; intravenous drug use; tattooing; acupuncture therapy; ear piercing; and major or minor surgery). Cleaning and Disinfection Method The instruments used for the known HCV carriers were not handled differently from those used for the HCV-negative patients; they were not removed from the general instrument pool, were disinfected in the same way as the others, and were then used promptly to perform endoscopy on the HCV-negative patients. Moreover, endoscopic procedures in known HCV carriers were not postponed at the end of the session but were performed according to the list of scheduled appointments. All units participating in this study adhered to the international guidelines for cleaning and disinfection practices in digestive endoscopy (6-10) and reprocessing endoscopic accessories (11); written protocols were available in each center. The staff involved in disinfection procedures consisted of trained nurses who were unaware of the ongoing study. At the end of the endoscopic procedure, the staff manually cleaned the instrument, including brushing the channels; each internal channel was flushed with detergent, rinsed with water, and blown through with air. The endoscopic units used 3 different automated washer-disinfectors (DSD-91E, Medivators, Minneapolis, Minnesota; Circlean MC-12, Shoei, Tokyo, Japan; and ETD2, Olympus Europe), but the reprocessing cycle was similar: 1) The units were immersed in 2% glutaraldehyde for 20 minutes, and internal channels were flushed with the same solution; 2) the units were rinsed internally and externally with drinking-quality water to remove all traces of disinfectant; and 3) the units were dried externally and each channel was flushed with air. Before the first endoscopy of each day, all endoscopes were disinfected in a washer-disinfector. After use, reusable biopsy forceps were immersed in enzymatic detergent solutions; next, they were cleaned first manually and then by a medical-grade ultrasonic cleaner. After rinsing and drying, the forceps were sterilized by autoclave at 134 C for at least 5 minutes. Finally, the sterilized devices were stored in sterile packaging in a closed cupboard where they were protected from dust, humidity, and temperature fluctuations. Laboratory Methods We tested for anti-HCV by using a third-generation enzyme immunoassay (Ortho HCV EIA-3, Ortho Diagnostic Systems, Raritan, New Jersey). Anti-HCV immunoreactivity was confirmed with a third-generation immunoblot assay (RIBA-3, Chiron Corp., Emeryville, California, and Ortho Diagnostic Systems). We measured HCV RNA by using PCR (Cobas Amplicor 2.0, Roche Diagnostic Systems, Branchburg, New Jersey); the sensitivity of this assay was 1000 copies/mL. Statistical Analysis We estimated person-years of observation and incidence rates of anti-HCV seroconversion for both cohorts. We used the difference between the incidence rates to compare the 2 cohorts. For the endoscopy cohort, we also measured the risk for anti-HCV seroconversion 6 months after the procedure, using the number of persons as denominators. Further analyses were limited to subgroups of the endoscopy cohort: 1) 6132 patients who underwent biopsy (biopsy cohort) and 2) 912 patients who underwent endoscopy later in the same day as and with the same instruments used in HCV-positive patients (at-risk cohort). Because we could not identify with certainty each patient in the at-risk cohort, we estimated that number with a rough but conservative approach. If we assume that each endoscope was used approximately 3 times during an ordinary endoscopic session and that at least 1 HCV carrier would be seen at

Combined evaluation of total and functional liver plasma flows and intrahepatic shunting

A diagnostic protocol was studied, designed to evaluate the main parameters of liver circulation in man. A water solution ofd-sorbitol (S) and indocyanine green (ICG) was infused intravenously in six controls and nine cirrhotics. Steady-state renal and hepatic S clearances as well as hepatic ICG clearance were calculated. In controls the values (mean ±sd) of the independent measurements of S and ICG hepatic clearance were 978±107 and 519±142 ml/min, respectively, while in cirrhotic patients they were 554±238 and 231±90 ml/min. Owing to the kinetic properties of S, its hepatic clearance may be regarded as a measure of functional liver plasma flow (FLPF). The total liver plasma flow (TLPF) values (mean±sd), calculated according to Ficks principle, were 1091±157 ml/min (S method) and 1033±153 ml/min (ICG method) in controls, and 1251±554 and 1284±677 ml/min in cirrhotics. In controls, FLPF was found to be very close to TLPF. In cirrhotic patients the difference between TLPF and FLPF (ranging from 169 to 2093 ml/min when measured by S method) was considered as an approximate estimate of intrahepatic shunting. The procedure is safe and simple and may add a new dimension to the investigation of hepatic circulation.

The use of probiotics in healthy volunteers with evacuation disorders and hard stools: A double-blind, randomized, placebo-controlled study

Background Evacuation disorders and hard stools are common in industrialized countries, affecting on average 12% to 17% of the adult healthy population at any age. Dietary supplementation with probiotic microorganisms may be useful in reducing the disorder. Methods We performed a double-blind, randomized, placebo-controlled study to evaluate the effectiveness of 2 different probiotic blends, either mixed Lactobacillus plantarum LP01 (LMG P-21021) and Bifidobacterium breve BR03 (DSM 16604) or Bifidobacterium animalis subspecies lactis BS01 (LMG P-21384), in the management of evacuation disorders and intestinal discomfort. In a period of 5 years (2003 to 2008), the study involved 300 healthy volunteers (151 males and 149 females; age 24 to 71 y) with evacuation disorders and hard stools. In particular, subjects were divided into 3 groups: 80 subjects in the group A received placebo, 110 subjects in the group B received mixed L. plantarum LP01 and B. breve BR03 (2.5×109 colony-forming units/d of each strain), and 110 subjects in the group C received B. animalis subsp. lactis BS01 (5×109 colony-forming units/d) for 30 days. At the beginning of the observational study, the healthy status of volunteers was evaluated by a complete, laboratory and ultrasound study of the abdomen. The physical examination was repeated after 15 and 30 days. In particular, the main troubles typically associated with evacuation disorders and hard stools as well as abdominal bloating were considered as parameters of interest. Exclusion criteria were items of gastrointestinal diseases and antibiotics intake. Results Subjects treated with the mixed probiotic strains L. plantarum LP01 and B. breve BR03 or B. animalis subsp. lactis BS01 reported a significant improvement in the number of weekly bowel movements and in the main troubles associated with evacuations, particularly consistency of feces and ease of expulsion. Discomfort items such as abdominal bloating and anal itching, burning, or pain also registered a relevant improvement in the active groups receiving probiotics. Conclusions The intake of an effective amount of mixed L. plantarum LP01 and B. breve BR03 or B. animalis subsp. lactis BS01 for 30 days is able to significantly relieve the evacuation disorders and hard stools, thus providing a useful tool for the management of such condition, which is particularly widespread in industrialized countries at any age.

Is microencapsulation the future of probiotic preparations? The increased efficacy of gastro-protected probiotics

In a recent publication we assessed the kinetics of intestinal colonization by microencapsulated probiotic bacteria in comparison with the same strains given in an uncoated form. It’s well known, in fact, that microencapsulation of probiotics with specific materials is able to confer a significant resistance to gastric juice, thus protecting the cells during the gastric and duodenal transit and enhancing the probiotic efficacy of any supplementation. In any case, this was the first study reporting the fecal amounts of probiotics administered in a coated, protected form compared with traditional, uncoated ones. Here we discuss additional in vitro data of resistance of the same bacteria to gastric juice, human bile and pancreatic secretion and correlate them with the results of in vivo gut colonization. .

In vitro sensitivity of probiotics to human pancreatic juice.

Background The resistance of gut flora with probiotic activity to pancreatic juice is usually tested with artificial pancreatic fluid. Previous studies evaluated the sensitivity of diverse probiotics to human gastric and biliary secretion; none tested the resistance of probiotics to human pancreatic juice. As most bacteria sensitive to artificial pancreatic fluid in vitro have a high rate of isolation from feces, the resistance to human pancreatic juice could be higher. Aim The aim of this study was to compare the sensitivity of different strains of probiotics to artificial and human pancreatic juice. Materials and Methods The viability of 8 strains of Lactobacillus and 4 strains of Bifidobacterium was tested with standard artificial and human pancreatic juice withdrawn from 16 patients during endoscopic retrograde cholangiopancreatography procedure. The mortality rate (%) of various bacteria was measured after 5, 30, and 60 minutes contact time. The results were normalized for mortality rate induced by hypotonic condition and time exposure to 37°C. Results When incubated with artificial and human pancreatic juice, the mortality rate of all strains of Lactobacillus at 5, 30, and 60 minutes was 10.1 versus 7.6, 20.5 versus 19.7, and 28.6 versus 29.8, respectively; whereas the mortality rate of all strains of Bifidobacterium was 8.0 versus 9.2, 33.3 versus 28.9, and 42.2 versus 44.4, respectively. Conclusions All the tested strains were sensitive to artificial and human pancreatic juice depending on time contact. Bifidobacterium strains seem to be more sensitive than Lactobacillus strains in particular at higher time contact. There is no significant difference between sensitivity to simulated and human pancreatic juice. For this reason, probiotics activity may be tested with artificial pancreatic fluid using a standardized, easier, and less costly procedure.

The innovative potential of Lactobacillus rhamnosus LR06, Lactobacillus pentosus LPS01, Lactobacillus plantarum LP01, and Lactobacillus delbrueckii Subsp. delbrueckii LDD01 to restore the "gastric barrier effect" in patients chronically treated with PPI: A pilot study

Background: Gastroesophageal reflux disease is a very widespread condition. In Europe, it is estimated that about 175 million people suffer from this disease and have to chronically take drugs to increase gastric pH. The proton pump inhibitors (PPIs) such as omeprazole, lansoprazole, and esomeprazole are the most widely used drug typology in this regard. However, the inhibition of normal gastric acid secretion has important side effects, the most important being bacterial overgrowth in the stomach and duodenum with a concentration of >105 viable cells/mL. As a major consequence of this, many harmful or even pathogenic bacteria contained in some foods could survive the gastric transit and colonize either the stomach itself, the duodenum, or the gut, where they could establish acute and even chronic infections with unavoidable consequences for the host’s health. In other words, the “gastric barrier effect” is strongly reduced or even disrupted. To date, there are no real strategies to deal with this widespread, although still relatively little known, problem. The aim of this study was to confirm the gastric bacterial overgrowth in long-term PPI consumers and to assess the efficacy of some probiotic bacteria, belonging to both genera Lactobacillus and Bifidobacterium, in the reduction of gastric and duodenal bacterial overgrowth, therefore partially restoring the gastric barrier effect against foodborne pathogenic bacteria. Methods: For this purpose, probiotics with a strong demonstrated inhibitory activity on gram-negative bacteria, such as Escherichia coli, were tested in a human intervention trial involving a total of 30 subjects treated with PPIs for either 3 to 12 consecutive months (short-term) or >12 consecutive months (long-term). An additional 10 subjects not taking PPIs were enrolled and used as a control group representing the general population. Four selected probiotics Probiotical SpA (Novara, Italy), namely Lactobacillus rhamnosus LR06 (DSM 21981), Lactobacillus pentosus LPS01 (DSM 21980), Lactobacillus plantarum LP01 (LMG P-21021), and Lactobacillus delbrueckii subsp. delbrueckii LDD01 (DSM 22106) were administered for 10 days to 10 subjects treated with PPIs for >12 months (group B). In the 60 mg formulation, N-acetylcysteine was included as well in light of its well-known mechanical effects on bacterial biofilms. Gastroscopies were performed at the beginning of the study (d0) in all the groups (A, B, C, and D) and after 10 days (d10) in group B only; that is, at the end of probiotics intake. The total viable cells and total Lactobacillus were quantified in gastric juice and duodenal brushing material from all subjects. The results were compared among all the groups and with the control subjects (group D) to confirm the bacterial overgrowth. A comparison was made also between d0 and d10 in group B to quantify the efficacy of the 4 probiotics administered for 10 days. Fecal samples were collected from all groups at d0, including subjects not treated with PPIs, and in group B only at d10. Specific bacterial classes, namely enterococci, total coliforms, E. coli, molds, and yeasts were quantified in all fecal specimens. Results: The results collected confirmed the strong bacterial overgrowth in the stomach and duodenum of people treated with PPIs compared with subjects with a normal intragastric acidity. It is also worth noting that the bacterial cell counts in subjects who underwent a long-term treatment with a PPI were greater than the results from subjects taking these drugs for 3 to 12 months. The intake of 4 specific probiotic strains with a marked antagonistic activity towards 5 E. coli bacteria, including the enterohaemorrhagic O157:H7 strain, and an effective amount of N-acetylcysteine (NAC) was able to significantly reduce bacterial overgrowth in long-term PPI-treated subjects. Total lactobacilli represented the major percentage of bacterial counts, thus demonstrating the ability of such bacteria to colonize the stomach and the duodenum, at least temporarily, and to consequently restore the gastric barrier effect. A significant decrease in fecal enterococci, total coliforms, E. coli, molds, and yeasts in subjects treated with PPIs was recorded at the end of probiotics supplementation (d10) compared with baseline (d0) in group B. This is a further confirmation of the barrier effect also exerted at the stomach level. Conclusions: PPIs are the most widely sold and used drugs in the world. However, the chronic use of these pharmacological molecules exposes the subject to the risk of foodborne infections as most pathogens are able to survive the gastric transit in a condition of significantly decreased acidity.

Non-invasive evaluation of portal-systemic shunting in man by d-sorbitol bioavailability

Portal-systemic shunting is an important circulatory abnormality in patients with cirrhosis. This study explores the potential of the natural polyol D-sorbitol as test compound for non-invasive assessment of shunting. Ten normal subjects, 10 patients with cirrhosis and 12 cirrhotics with surgical portacaval shunts were studied after oral and intravenous administration of a 2 g dose of sorbitol. As measured by the H2 breath test, removal from the intestinal lumen was complete in both groups. Bioavailability of sorbitol, calculated as ratio of the areas under the plasma concentration/time curve after p.o. and i.v. administration, was zero in normal subjects, 0.29 +/- 0.15 in cirrhotic patients, and 0.38 +/- 0.11 in patients with portacaval shunts. Calculation of bioavailability on the basis of urinary outputs of sorbitol gave similar results. It is concluded that the bioavailability of sorbitol reflects portal-systemic shunting, although the relatively low figures suggest some degree of sorbitol metabolism by enterocytes.

The "Prometeo" study: online collection of clinical data and outcome of Italian patients with acute nonvariceal upper gastrointestinal bleeding.

Goals: To implement an online, prospective collection of clinical data and outcome of patients with acute nonvariceal upper gastrointestinal bleeding (UGIB) in Italy (“Prometeo” study). Background: Epidemiology, clinical features, and outcomes of nonvariceal UGIB are mainly known by retrospective studies and are probably changing. Study: Data were collected by 13 Gastrointestinal Units in Italy from June 2006 to June 2007 (phase 1) and from December 2008 to December 2009 (phase 2): an interim analysis of data was performed between the 2 phases to optimize the online database. All the patients consecutively admitted for acute nonvariceal UGIB were enrolled. Demographic and clinical data were collected, a diagnostic endoscopy performed, with endoscopic hemostasis if indicated. Results: One thousand four hundred thirteen patients (M=932, mean age±SD=66.5±15.8; F=481, mean age±SD=74.2±14.6) were enrolled. Comorbidities were present in 83%. 52.4% were treated with acetyl salicylic acid or other nonsteroidal anti-inflammatory drugs (NSAIDs): only 13.9% had an effective gastroprotection. Previous episodes of UGIB were present in 13.3%. Transfusion were needed in 43.9%. Shock was present in 9.3%. Endoscopic diagnosis was made in 93.2%: peptic lesions were the main cause of bleeding (duodenal ulcer 36.2%, gastric ulcer 29.6%, gastric/duodenal erosions 10.9%). At endoscopy, Helicobacter pylori was searched in 37.2%, and found positive in 51.3% of tested cases. Early rebleeding was observed in 5.4%: surgery was required in 14.3% of them. Bleeding-related death occurred in 4.0%: at multivariate analysis, the risk of death was correlated with female sex [odds ratio (OR=2.19, P=0.0089)], presence of neoplasia (OR=2.70, P=0.0057) or multiple comorbidities (OR=5.04, P=0.0280), shock at admission (OR=4.55, P=0.0001), and early rebleeding (OR=1.47, P=0.004). Conclusions: Prometeo database has provided an up-to-date picture of acute nonvariceal UGIB in Italy: patients are elderly, predominantly males, and with important comorbidities. Gastroprotection is underutilized during NSAIDs treatment. With respect to previous studies, Prometeo shows a higher incidence of low-dose acetyl salicylic acid use and comorbidities, whereas no significant difference were found in other items (etiology of bleeding, NSAIDs use, need for endoscopic hemostasis, incidence of rebleeding, and overall mortality).

A multistep cytological approach for patients with jaundice and biliary strictures of indeterminate origin

Aims Fluorescence in situ hybridisation (FISH) increases the sensitivity for detecting pancreatobiliary tract cancer over routine cytology. In this study, diagnostic accuracy and costs of cytology and FISH in detecting cancer in patients with jaundice with biliary strictures were assessed. Methods Brushing specimens from 109 patients with jaundice were obtained during endoscopic retrograde cholangiopancreatography and examined by cytology and FISH. The specimens were considered FISH-positive for malignancy if at least five polysomic cells or 10 cells with homozygous or heterozygous 9p21/p16 deletion were detected. Definitive diagnosis of the stricture as benign or malignant relied on surgical pathology (45 cases) or clinical-radiological follow-up >18 months (64 cases). We calculated costs of cytology and FISH based on the reimbursement from the Piedmont region, Italy (respectively, €33 and €750). Results Ninety of 109 patients had evidence of malignancy (44 pancreatic carcinomas, 36 cholangiocarcinomas, 5 gallbladder carcinomas, 5 other cancers), while 19 had benign strictures. Routine cytology showed 42% sensitivity, but 100% specificity for the diagnosis of malignancy, while FISH-polysomy showed 70% sensitivity with 100% specificity and FISH-polysomy plus homozygous or heterozygous 9p21/p16 deletion showed 76% sensitivity with 100% specificity. The cost per additional correct diagnosis of cancer obtained by FISH, in comparison with cytology, was €1775 using a sequential cytological approach (ie, performing FISH only in patients with negative or indeterminate cytology). Conclusions FISH should be recommended as the second step in detecting cancer in patients with jaundice with pancreatobiliary tract strictures and cytology negative or indeterminate for malignancy.