Autopercepción de la imagen corporal en estudiantes universitarios de Chile y Panamá / Body image perception in university students from Chile and Panama

Objetivo: Determinar la concordancia entre el estado nutricional y la percepción de la imagen corporal en universitarios chilenos y panameños. Métodos: Estudio observacional transversal en 792 estudiantes de ambos sexos, de 18 a 30 años de edad de la Universidad Santo Tomás de Vina del Mar en Chile, la Universidad Autónoma de Chile y la Universidad Latinoamericana de Ciencia y Tecnología (ULACIT) de Panamá. Se determinó el IMC real y el IMC percibido y luego éstos fueron comparados. Resultados: De 792 estudiantes evaluados, 480 eran de Chile (51,4% mujeres) y 312 de Panamá (56,7% mujeres). El IMC percibido por los chilenos fue mayor al real, en hombres y mujeres (60%); en cambio, los panameños presentan una sobreestimación menor al 50% en ambos sexos. En ambos países y para ambos sexos el porcentaje de estudiantes que se perciben igual es similar (20%). Hay concordancia diagnóstica entre el IMC y la imagen corporal en la mitad de los alumnos chilenos y en un tercio de las alumnas, pero la mejor concordancia se observó en personas con sobrepeso (69% y 90,5% respectivamente). En los panameños se observó concordancia diagnóstica de 40% en varones y 19,2% en mujeres; la mejor concordancia se observó en hombres con sobrepeso (83%). Conclusión: En los estudiantes con normopeso de ambos países existe una tendencia a sobreestimar el peso, mientras que por el contrario los obesos tienden a subestimarlo.

Objective: To determine the association between nutritional status and body image perception among university students from Chile and Panamá. Methods: An observational cross-sectional study was conducted among 792 students of both sexes, ages 18 to 30 years, from University Santo Tomás of Viña del Mar, Chile, University Autónoma of Chile and University Latinoamericana de Ciencia y Tecnología (ULACIT) of Panama. We determined the real BMI and the perceived BMI. Afterward, the two were compared to each other. Results: In a sample of 792 students, 480 from Chile (51,4% females) and 312 from Panamá (56,7% females), the perceived BMI is higher than the real one in the Chileans, men and women (60%). Instead, in Panamá the over-estimation is less than 50% in both sexes. The percentage is the same in both countries and both sexes that perceive themselves as they really are (20%). There is concordance between BMI and body image in half of Chilean men and a third of Chilean women. However, a better concordance is observed in overweight students (69% in men, 90,5% in women). In Panamanians, there is a concordance observed of40% in men and 19,9% in women. Conclusion: There is a tendency for overestimation in normal weight students of both countries. However the obese students tended to underestimate their weight.

Pathological and mammographic prognostic factors for screen detected cancers in a multi-centre randomised, controlled trial of mammographic screening in women from age 40 to 48 years.

AIM: To assess pathological and radiological prognostic factors for cancers detected by screening within a multi-centre RCT trial of mammographic screening of younger women. METHOD: The survival of 232 women with screen detected invasive cancer was ascertained. Data on invasive cancer size, histological grade, nodal status, vascular invasion, mammographic spiculation, comedo calcification and mammographic background were assessed. Kaplan-Meier and Cox proportional hazards methods were used to examine survival. RESULTS: Univariate analysis indicated that women with cancers with the following features had poorer survival; ≥ 30 mm, histologically grade 3, heavily node positive (4 or more positive nodes), vascular invasion positive and displaying mammographic comedo calcification. In a multivariate model survival remained poorer in women with four or more nodes positive (HR 8.36, 95% CI 2.31, 30.17) and in those with comedo calcification (HR 3.00,95% CI 1.13, 7.99). CONCLUSION: Nodal status and the presence of mammographic comedo calcification have independent prognostic significance in young women with screen detected cancer.

Hepatic encephalopathy after transjugular intrahepatic portosystemic shunt insertion: a decade of experience.

BACKGROUND: Since its introduction, transjugular intrahepatic portosystemic shunt (TIPS) has been extensively used for treatment of portal hypertension. We report a decade of experience with particular emphasis on characterizing post-TIPS hepatic encephalopathy (HE). AIM: To determine the frequency of clinically evident or minimal HE post-TIPS, identify predisposing factors and determine the impact of minimal HE on quality of life. DESIGN: Prospective data collection and retrospective case notes analysis. METHODS: Of 197 patients referred for TIPS insertion, 136 patients who survived the procedure by more than 4 weeks were available for assessment. Data collected at TIPS insertion was supplemented by case note analysis. Psychometric testing was performed and health profile questionnaires administered on patients still attending. RESULTS: Most patients had alcoholic liver disease (62.4%) and bleeding varices unresponsive to endoscopic therapy (86%). Clinically evident post-TIPS HE developed in 34.5% of patients, was of similar frequency in the groups treated with polytetrafluoroethylene covered and uncovered stents, and the only significant predictor was pre-TIPS HE. Post-TIPS HE necessitating liver transplant or contributing to death occurred in only 14 (10.3%) patients. Minimal encephalopathy (abnormal psychometry) was present in 49% of patients at 26 (3-123) months after TIPS but this frequency was similar in a cohort of cirrhotics being assessed for liver transplant. However, patients with abnormal psychometry had significantly lower quality of life scores than those with normal psychometry. CONCLUSION: Although, HE is relatively common after TIPS insertion, with careful selection of patients it is usually short-lived and easily managed. Minimal HE is no more prevalent than expected in a cirrhotic population without TIPS.

Radiological and pathological findings of interval cancers in a multi-centre, randomized, controlled trial of mammographic screening in women from age 40-41 years.

AIM: The aim of this study was to analyse the radiographic findings of the screening mammograms of women with interval cancer who participated in a multi-centre, randomized, controlled trial of mammographic screening in women from age 40-48 years. MATERIALS AND METHODS: The screening and diagnostic mammograms of 208 women with interval cancers were reviewed. Abnormalities were classified as malignant, subtle and non-specific. RESULTS: Eighty-seven (42%) of women had true, 66 (32%) occult and 55 (26%) false-negative interval cancers. The features most frequently missed or misinterpreted were granular microcalcification (38%), asymmetric density (27%) and distortion (22%). Thirty-seven percent of abnormal previous screens were classified as malignant, 39% subtle change and 21% as non-specific. Granular calcifications were significantly more common on the diagnostic mammograms of false-negative interval cancers than those of true interval cancers (28 versus 14%, p=0.04). Occult interval cancers were more likely to be <10 mm and <15 mm in invasive pathological size than other interval cancers (p=0.03 and 0.005, respectively). True interval cancers were more likely to be histologically grade 3 than other interval cancers (p=0.04). Women who developed true and false-negative interval cancers had similar background patterns, but women with occult cancers had a higher proportion of dense patterns (p<0.05). CONCLUSION: Interval cancers in a young screening population have a high proportion of occult lesions that are small and occur in dense background patterns. The proportion of interval cancers that are false negative is similar that seen in older populations and granular microcalcification is the commonest missed mammographic feature.

Radiological findings of screen-detected cancers in a multi-centre randomized, controlled trial of mammographic screening in women from age 40 to 48 years.

AIM: To elucidate the mammographic findings of screen-detected cancers in women screened between 40-48 years, and to establish the frequency and nature of abnormal findings on previous mammograms in women with screen-detected cancers. METHODS: A radiology review panel consisting of three experienced breast radiologists viewed the screening mammograms in chronological order, confirming that any abnormalities detected corresponded to the cancers detected at later screens. An analysis correlating mammographic features with median invasive size and the proportion measuring less than 10 and 15mm was performed. RESULTS: Two hundred and thirty-two women had screen-detected invasive cancers with mammograms available for review. The most frequent features seen at diagnosis were spiculate mass, ill-defined mass, granular calcification, deformity and comedo calcification. Thirty-four percent of mammograms showed calcification. The mammographic sign associated with smallest median size was calcification. Calcification was also the mammographic abnormality most frequently associated with cancers <10mm in size. In total there were 147 abnormal previous screens of 87 women. The most commonly missed features were granular microcalcification, deformity and ill-defined mass. Of the missed abnormalities 20% were classified as malignant, 43% as subtle change and 32% as non-specific. CONCLUSION: Compared with older women, screen-detected cancer in younger women more commonly manifests as calcification and less frequently a spiculate mass. Calcification is the sign most frequently associated with invasive cancers <10mm in size. Calcification and deformity are the signs most frequently seen on the previous mammograms of women with screen-detected cancer.

Detection assessment and monitoring of hepatic fibrosis: biochemistry or biopsy?

Hepatic fibrosis is an important consequence of inflammatory disorders affecting the liver, and ultimately progresses to cirrhosis. Here we explore methods for the detection and monitoring of hepatic fibrosis, particularly in hepatitis C, alcoholic liver disease, non-alcoholic fatty liver disease and during methotrexate therapy, in all of which progressive fibrosis can develop over a number of years in a minority of patients. Liver biopsy currently remains the gold standard to assess fibrosis. However, it has several limitations, including manpower issues, cost, risk of patient injury, including mortality and morbidity, observer variability and sampling variation. Several non-invasive diagnostic tests for fibrosis and cirrhosis have therefore been evaluated. The usefulness of a laboratory test for screening for a pathological abnormality such as fibrosis is critically dependent on the prevalence of the pathology in the population under investigation. When the prevalence is expected to be low, screening tests should have a high negative predictive value so that large numbers of patients can be spared the next diagnostic step, namely liver biopsy. For the moment, clinical chemistry laboratories should offer the aspartate aminotransferase alanine aminotransferase ratio, AST/platelet ratio and the Rosenberg fibrosis index as part of their routine service for monitoring the development of hepatic fibrosis.

Managing errors in radiology: a working model.

AIM: To develop a practical mechanism for reviewing reporting discrepancies as addressed in the Royal College of Radiologists publication "To err is human. The case for review of reporting discrepancies". MATERIALS AND METHODS: A regular meeting was developed, and has evolved, within the department to review discrepancies. Standard forms were devised for submission of cases as well as recording and classification of discrepancies. This has resulted in availability of figures that can be audited annually. RESULTS: Eighty-one cases involving error were reviewed over a 12-month period. Seven further cases flagged as discrepancies were not identified on peer review. Twenty-four reports were amended subsequent to the meeting. Nineteen additional cases were brought to the meeting as illustrative of teaching points or for discussion. CONCLUSION: We have evolved a successful process of reviewing reporting errors, which enjoys the confidence and support of all clinical radiologists, and is perceived as a method of improving patient care through an increasing awareness of lapses in performance.

Monitoring methotrexate-induced hepatic fibrosis in patients with psoriasis: are serial liver biopsies justified?

BACKGROUND: Reports that up to 26% of subjects with psoriasis develop cirrhosis have led to a recommendation of serial liver biopsies after each cumulative dose of 1500 mg of methotrexate. AIM: To evaluate the progression of liver injury in patients with psoriasis and the impact of monitoring by liver biopsy on their management. METHODS: One hundred and twenty-one liver biopsies from 66 subjects (aged 11-79 years) with psoriasis, receiving a median cumulative dose of 3206 mg of methotrexate over a period of 280.5 weeks, were evaluated. RESULTS: The assessment of advanced fibrosis according to the Ishak system (>or= 4) correlated perfectly with that of the Scheuer system (>or= 3) and poorly with that of the Roenigk scale (>or= 3b) (r2 = 1.0 and 0.31, respectively). Two of 24 pre-treatment biopsies showed advanced fibrosis and both subjects were heavy drinkers. The cumulative probabilities of advanced fibrosis (Ishak >or= 4) were 0%, 2.6%, 2.6%, 8.2% and 8.2% at cumulative doses of 1500, 3000, 4500, 5000 and 6000 mg, respectively. None of the subjects developed cirrhosis during follow-up or discontinued therapy on the basis of liver biopsy findings. CONCLUSIONS: Advanced hepatic fibrosis with low-dose methotrexate therapy is much less frequent than previously reported. Pre-treatment or monitoring liver biopsies in accordance with the current guidelines have little impact on patient management.

Amino acid challenge in patients with cirrhosis: a model for the assessment of treatments for hepatic encephalopathy.

BACKGROUND/AIMS: To mimic episodic hepatic encephalopathy after gastrointestinal bleeding under controlled conditions, cirrhotic patients were challenged with an amino acid mixture of comparable composition to haemoglobin. METHODS: Basal EEG, psychometric score (HE test), reaction times and venous blood ammonia were recorded. Following a 54 or 108 gm oral amino acid challenge, blood ammonia levels and EEG were recorded at 30-min intervals, and psychometric testing was repeated at 180 min. Ten controls (57 +/- 2) and 31 cirrhotics (52 +/- 2) of which 21 were Child's grade A or B and 10 grade C underwent the challenge. Nine had a transjugular intrahepatic porta-systemic shunt in situ. RESULTS: Seventeen patients had abnormal baseline HE scores. Basal blood ammonia and reaction time A were significantly greater in patients (52 +/- 5 micromol/l and 478 +/- 20 ms, respectively) than controls (19 +/- 2 micromol/l and 372 +/- 14 ms) (P < 0.001). Following the challenge, in patients with advanced liver disease (Child's grade B and C) the slowing of reaction time A (+85 +/- 38 and +71 +/- 31 ms, respectively; P < 0.03) and EEG (ratio of slow to fast wave activity +0.31 +/- 0.12 and +0.58 +/- 0.19; P < 0.02) were significantly greater than in controls (-3.3 +/- 8 ms and 0.00 +/- 0.03, respectively). Patients with an abnormal basal HE score had the most pronounced changes (reaction time A +110 +/- 39 ms, P < 0.01, EEG +0.52 +/- 13, P < 0.01, respectively). The change in EEG ratio correlated with the dose of amino acid administered (r = 0.96; P < 0.008). CONCLUSION: The amino acid challenge constitutes a reproducible human model of episodic, Type C hepatic encephalopathy unaffected by the complications usually encountered in clinical practice.

Quantitative assessment of fibrosis and steatosis in liver biopsies from patients with chronic hepatitis C.

BACKGROUND: Hepatic fibrosis is one of the main consequences of liver disease. Both fibrosis and steatosis may be seen in some patients with chronic hepatitis C and alcoholic liver disease (ALD). AIMS: To quantitate fibrosis and steatosis by stereological and morphometric techniques in patients with chronic hepatitis C and compare the results with a control group of patients with ALD. In addition, to correlate the quantitative features of fibrosis with the Ishak modified histological score. MATERIALS AND METHODS: Needle liver biopsies from 86 patients with chronic hepatitis C and from 32 patients with alcoholic liver disease (disease controls) were analysed by stereological and morphometric analyses using the Prodit 5.2 system. Haematoxylin and eosin and Picro-Mallory stained sections were used. The area fractions (A(A)) of fibrosis, steatosis, parenchyma, and other structures (bile duct and central vein areas) were assessed by stereological method. The mean diameters of fat globules were determined by morphometric analysis. RESULTS: Significant differences were found in the A(A) of fibrosis, including fibrosis within portal tract areas, between chronic hepatitis C patients and those with ALD (mean (SD): 19.14 (10.59) v 15.97 (12.51)). Portal and periportal (zone 1) fibrosis was significantly higher (p = 0.00004) in patients with chronic hepatitis C compared with the control group (mean (SD): 9.04 (6.37) v 3.59 (3.16)). Pericentral fibrosis (zone 3) occurred in both groups but was significantly more pronounced in patients with ALD. These results correlate well with the modified Ishak scoring system. However, in patients with cirrhosis (stage 6) with chronic hepatitis C the A(A) of fibrosis varied between 20% and 74%. The diameter of fat globules was significantly lower in patients with hepatitis C (p = 0.00002) than the ALD group (mean (SD): 14.44 (3.45) v 18.4 (3.32)). Microglobules were more frequent in patients with chronic hepatitis C than in patients with ALD. In patients with chronic hepatitis C, the fat globules had a zonal distribution in comparison with pan steatosis in ALD. CONCLUSION: Quantitative, stereological techniques are simple and reliable for evaluating hepatic fibrosis and steatosis in chronic hepatitis C. They are most useful for assessing the origin, location, and the stage of fibrosis. Stereology and morphometry are recommended for the quantitation of fibrosis and steatosis, particularly for the evaluation of new treatment strategies in patients with chronic hepatitis C.

Effect of L-ornithine-L-aspartate on patients with and without TIPS undergoing glutamine challenge: a double blind, placebo controlled trial.

BACKGROUND AND AIM: An oral glutamine load in cirrhotic patients awaiting liver transplantation was shown to cause a rise in blood ammonia and psychometric abnormalities which were reversed by hepatic transplantation. L-Ornithine-L-aspartate (LOLA) has been shown to reduce ammonia and improve psychometric function in patients with hepatic encephalopathy. The aim of the present study was to assess the effect of LOLA in healthy patients with cirrhosis and no evidence of clinical encephalopathy after challenging the central nervous system by administration of oral glutamine. PATIENTS AND METHODS: Eight cirrhotics (Child's B or C) without transjugular intrahepatic portosystemic shunts (TIPS) and seven with TIPS underwent two oral glutamine (20 g) challenges, receiving LOLA (5 g intravenously) on one occasion and placebo on the other in random order. Psychometric tests, including choice reaction time (CRT) and number connection test, were performed before and after glutamine, together with electroencephalography and blood ammonia. RESULTS: Mean basal ammonia was 27 (SEM 5) micromol/l in non-TIPS and 76 (10) micromol/l in TIPS patients (p<0.05). Basal CRT 2 was 0.643 (0.033) s in non-TIPS and 0.825 (0.076) s in TIPS patients (p<0.02). In non-TIPS patients, ammonia increased to 36 (10) micromol/l when LOLA was administered and to 62 (13) micromol/l with placebo (p<0.02). There was no alteration in psychometric function in non-TIPS patients after glutamine when LOLA was given but when placebo was given, glutamine caused prolongation of CRT (p=0.02). Glutamine did not affect psychometric function in TIPS patients with or without LOLA. CONCLUSION: This study showed that LOLA ameliorated the deleterious psychometric effects of glutamine in Child's grade B and C patients with cirrhosis without TIPS and supports its use in clinical practice in hepatic encephalopathy.

Do gastric and oesophageal varices bleed at different portal pressures and is TIPS an effective treatment?

BACKGROUND: It has been suggested that gastric varices bleed at lower portosystemic pressure gradients (PSPG) than oesophageal varices and that transjugular intrahepatic portosystemic shunt (TIPS) is a particularly effective treatment in these patients. AIMS: This study was undertaken to assess the difference in PSPG measured at the time of TIPS insertion between patients bleeding from gastric and those bleeding from oesophageal varices. Rebleeding and mortality rates between the two groups were also compared. PATIENTS AND METHODS: In a five year period, 64 patients (36 males and 28 females) undergoing TIPS for acute variceal bleeding had PSPG measured at the time of TIPS insertion. 12 patients underwent TIPS for gastric variceal haemorrhage (GVH) and 52 for oesophageal variceal haemorrhage (OVH). The median age was 53 years and 40/64 patients (63%) had alcoholic liver disease. The median Child's Pugh score was 8 for GVH and 9 for OVH patients. Median follow up was 75 weeks. RESULTS: There was no significant difference in median PSPG between patients with GVH, 21 mmHg (range 15-30 mmHg) and OVH, 22 mmHg (range 12-45 mmHg). Following TIPS, PSPG was 8.5 mmHg (range 3-11 mmHg) and 9 mmHg (range 4-20 mmHg) in GVH and OVH patients respectively. Rebleeding occurred in 2/12 (16%) GVH patients and 12/52 (23%) OVH patients (p= 1.0). Mortality during follow up was 25% (4/12) in the GVH and 25% (13/52) in the OVH patients. CONCLUSION: In this study, there was no difference between the pressures at which gastric and oesophageal varices bled. Rebleeding and mortality rates were similar in the two groups. TIPS is equally effective in the treatment of both oesophageal and gastric variceal haemorrhage.

Hepatitis C virus genotypes in a cohort of Middle Eastern patients.

BACKGROUND: The epidemiology of hepatitis C virus infection has been well characterized in Western Europe, North America and Japan. Less is known about it in other regions of the world. In order to fully understand the relationship between host and virus, it is important to study the effect of virus infection in all regions of the world. In this report, we have analyzed patients from the United Arab Emirates, Egypt and Jordan. DESIGN AND METHODS: Serum from 81 Middle Eastern HCV ELISA-2-positive patients was analyzed for the presence of HCV RNA by PCR. RNA-positive patients were genotyped by selective hybridization of amplicons to HCV genotype-specific oligonucleotides (InnoLipa2, Innogenetics, Belgium). Where possible, data was also obtained on racial origin, liver histology, serum ALT, prothrombin time, albumin, and risk factors for infection. RESULTS: Sixty-five of 81 patients were HCV RNA-positive. A higher proportion of Middle Eastern patients were genotype 4 compared to equivalent studies from Western Europe, USA and Japan. However, the most common genotype was 1a. No significant difference in genotype was found between patients with chronic hepatitis and patients with cirrhosis. CONCLUSIONS: Eight of 65 (12%) patients were genotype 4, but the most common genotype was 1a, a âWesternâ genotype (24/65, 37%). The mean age of cirrhotics was low compared to Western studies. This may be due to infection in early childhood or race-related host factors. Twelve of 65 patients (18%) were not classifiable for genotype using InnoLipa2. This may be due to multiple infecting genotypes in these patients, or unusual, non 13 HCV genotypes which cannot be classified by InnoLipa2.

Quantitative studies of testicular atrophy following portacaval shunt in rats.

To evaluate the differential effects of portacaval shunting (PCS) on the morphological changes that occur in humans with portal-systemic encephalopathy, male rats underwent either PCS (13) or sham operations (10). Normal adult rats (6) were used as controls. All animals were killed 5 to 7 weeks after the surgery. The wet weight of the testes was obtained. Hematoxylin-eosin (HE)-stained sections at 5-micrometers thickness were used for stereological analysis using an image analysis system. Apoptosis was assessed quantitatively in HE and in in situ end-labeling (ISEL)-stained slides, while mitotic activity and mast cell numbers were assessed in 20 high-power fields. There was a significant reduction in the testicular mass (664 mg) in PCS rats in comparison with sham (2,199 mg) and control (1,937 mg) rats (P <.00001). The thickness of germinal epithelium was significantly reduced in PCS rats (64 micrometers) compared with sham (126 micrometers) and control groups (108 micrometers). The number of tubules per square millimeter and the mean curvature were significantly increased in PCS rats (P <.00001). There was a 112-fold increase in apoptosis in PCS rats (112) in comparison with the control and sham-operation groups (1.2 and 0.7, respectively). Mitosis was significantly reduced in the PCS group (P =.0089), but mast cells were unchanged. The results suggest that PCS in the absence of liver dysfunction produces testicular atrophy by reduction in mitosis, maturation arrest, and increased apoptosis of the germinal epithelium. PCS may therefore be responsible for gonadal atrophy that occurs with advanced liver disease in humans.

Quantitative assessment of gastric atrophy using the syntactic structure analysis.

AIM: To assess the topographical relation between gastric glands, using the minimum spanning tree (MST), to derive both a model of neighbourhood and quantitative representation of the tissue's architecture, to assess the characteristic features of gastric atrophy, and to assess the grades of gastric atrophy. METHODS: Haematoxylin and eosin stained sections from corporal and antral biopsy specimens (n = 139) from normal patients and from patients with nonatrophic gastritis and atrophic gastritis of grades 1, 2, and 3 (Sydney system) were assessed by image analysis system (Prodit 5.2) and 11 syntactic structure features were derived. These included both line and connectivity features. RESULTS: Syntactic structure analysis was correlated with the semiquantitative grading system of gastric atrophy. The study showed significant reductions in the number of points and the length of MST in both body and antrum. The standard deviation of the length of MST was significantly increased in all grades of atrophy. The connectivity to two glands was the highest and most affected by the increased grade of atrophy. The reciprocal values of the Wiener, Randic, and Balaban indices showed significant changes in the volume of gland, abnormality in the shape of glands, and changes in irregularity and branching of the glands in both types of gastric mucosa. There was a complete separation in the MST, connectivity, and index values between low grade and high grade gastric atrophy. CONCLUSIONS: (1) Gastric atrophy was characterised by loss of the gland, variation in the volume, reduction in the neighbourhood, irregularity in spacing, and abnormality in the shape of the glands. (2) Syntactic structure analysis significantly differentiated minor changes in gastric gland (low grade atrophy) from high grade atrophy of clinical significance. (3) Syntactic structure analysis is a simple, fast, and highly reproducible technique and appears a promising method for quantitative assessment of atrophy.

Transjugular intrahepatic portosystemic shunt: a limited role in refractory ascites.

OBJECTIVE: To evaluate the role of the transjugular intrahepatic portosystemic shunt (TIPS) in the management of patients with refractory ascites. DESIGN: A retrospective study of 25 consecutive patients for whom refractory ascites was the primary indication for TIPS insertion. SETTING: Regional liver unit at Freeman Hospital, Newcastle upon Tyne, UK. PARTICIPANTS AND INTERVENTIONS: Twelve male and 13 female patients with a mean age of 58 years and mean Child-Pugh score of 10, treated with TIPS for refractory ascites between July 1992 and September 1995. MAIN OUTCOME MEASURES: Effect of TIPS on mortality, ascites and hospital admission rate. RESULTS: TIPS was successfully placed in all patients with a 59% mean reduction in portosystemic pressure gradient. Response rate was 68%, 48% and 33% at 1, 3 and 12 months, respectively. Mortality was 48% at 3 months and 67% at 12 months, being higher in those patients older than 60, those with renal impairment and those with higher Child-Pugh score. Amongst nine patients surviving long term (> 12 months) the mean time spent in hospital in the 3 months before TIPS was 35 days and in the year following TIPS 30 days. Patients who died (16 in total) spent a mean of 19 days in hospital before TIPS, 10 never leaving hospital, and 6 who were discharged spent a mean of 19 days post procedure in hospital (mean survival 84 days). CONCLUSION: TIPS has a limited role in the management of patients with refractory ascites. It is not an appropriate treatment where patients are older than 60, have renal impairment (creatinine > 200 mumol/l) or have a Child-Pugh score greater than 10.