FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W1923659905> ?p ?o ?g. }

• W1923659905 endingPage "1093" @default.
• W1923659905 startingPage "1091" @default.
• W1923659905 abstract "Gastro-oesophageal variceal bleeding is a severe and life-threatening complication of cirrhosis related to portal hypertension. The risk of bleeding is high for patients with large oesophageal varices or even small varices with red signs or in a decompensation setting. In these patients, the risk of bleeding can be reduced by about 50% using prophylaxis with β-blockers or band ligation (1-3). Therefore, authoritative guidelines recommend upper gastro-intestinal endoscopy for variceal detection in all patients with cirrhosis to be repeated after 1–3 years depending on the clinical situation and results of the first endoscopy (1-3). However, oesophageal varices occur in 30–40% of patients with compensated cirrhosis (1), and large oesophageal varices occur in only 9–36% of patients with cirrhosis (4). Consequently, most subjects who undergo upper endoscopy do not have varices and do not need prophylaxis. Moreover, one can guess that this proportion will grow because noninvasive screening of cirrhosis is widespread, where patients are diagnosed with a probable cirrhosis by serum tests or other noninvasive procedures at a very early, asymptomatic stage (5). Besides, upper endoscopy is an unpleasant, uncomfortable, invasive and costly procedure. In addition, the large number of endoscopies necessary for universal variceal screening implies a heavy burden for endoscopy units. Moreover, there is an interobserver variability in the interpretation of endoscopy results (5). Accordingly, several groups have devised predictors to discriminate patients with cirrhosis at a high risk for varices from those at a low risk, thus avoiding endoscopic screening in the latter category (4, 6-16). Most predictors are based on combinations of platelet level, aspartate aminotransferase (AST), alanine aminotransferase (ALT), AST/ALT ratio, age, spleen diameter, albumin, increased portal vein at ultrasonography, prothrombin activity, liver stiffness, ascites, Child–Pugh class and the FibroTest (4, 6-16). However, none of them has proved to be so accurate as to become a surrogate of upper endoscopy (1-3). In the July issue of Liver International, Kim et al. (17) report the results of a prospective, cross-sectional validation study of a simple index for prediction of oesophageal varices. This index, called ‘P2/MS’, is calculated by the formula [platelet count (109/L)]2/[monocyte fraction (%) × segmented neutrophil fraction (%)]. Therefore, only a complete blood count is required to compute this index. Kim and colleagues validated the P2/MS index, first proposed by Lee et al. (18), on a cohort of 318 consecutive patients with HBV-related cirrhosis. All patients underwent upper endoscopy. Two experienced endoscopists independently confirmed all endoscopic findings. The authors constructed receiver operator characteristic (ROC) curves for the P2/MS index as well as for other non-invasive scores used to detect varices and computed areas under the ROC curve (AUROC). AUROC correlates with the discriminating ability of a test; indeed, it ranges between 0.5 (no discrimination) and 1.0 (perfect discrimination). The AUROC for the P2/MS index was as high as 0.915 and 0.941 for the detection of all oesophageal varices and high-risk oesophageal varices, respectively, and was significantly higher than the AUROC of platelet count alone, namely 0.893 and 0.912 for all varices and for high-risk varices respectively. Of course, because platelet count is the numerator of the ratio, the lower the score, the higher the likelihood of oesophageal varices. Based on their results, the authors devised two cut-offs: values below 11 achieved a positive predictive value (PPV) of 94.0% (95% CI 86.5–98.0%), a specificity of 97.5% (95% CI 94.3–99.2%) and a positive likelihood ratio (PLR) of 26.44 (95% CI 11.02–63.43) in detecting high-risk varices; index values above 25 achieved a negative predictive value (NPV) of 94.4% (95% CI 90–97.3%), a sensitivity of 91.5% (95% CI 85–95.6%) and a negative likelihood ratio (NLR) of 0.1 (95% CI 0.06–0.18) for the detection of high-risk varices. Using these two cut-offs, the P2/MS score was able to determine reliably the likelihood of high-risk oesophageal varices in 262 patients (82.4%). The authors suggest that patients with a score <11 (26.1% of the total sample) should undergo endoscopy or begin β-blocker prophylaxis without endoscopy; patients with a score >25 (56.3% of the total sample) could be spared endoscopy and should undergo periodic evaluations of the score; and finally patients in the grey zone (11<P2/MS<25) (17.6% of the total sample) should undergo upper endoscopy. The authors provide an interesting comparison of P2/MS with other predictors. Based on AUROC, the diagnostic accuracy of the score was not significantly different from that of the spleen/platelet ratio (9) and the age/spleen/platelet ratio (but has the advantage of being much more simple), and was better than the other scores (7, 8, 10, 11). However, a comparison of scores in terms of sensitivity, specificity, NPV, PPV, PLR and NLR would have been more useful because these parameters have a major impact in clinical use than the AUROC. Further studies are required to compare other noninvasive scores with P2/MS. Moreover, this index has not been compared with other indexes, like the FibroTest, that have been largely validated in the setting of fibrosis. Several considerations arise from this paper: •A screening test should be very sensitive, i.e. it should detect all cases of high-risk varices. Regarding that point, the authors should be acknowledged for having performed analyses in order to assess the presence of large varices or varices with red signs. Consequently, a negative result should identify patients who, having a very low or null probability of high-risk varices, can safely avoid endoscopy. In other words, the NPV should be very close to 100%, provided that the proportion of patients affected by the threshold is not low, which is a major issue. The NPV of the P2/MS score, for a cut-off fixed at 25, is as high as 94.4%. This means that 5.6% of patients with a negative test have high-risk varices but should not undergo upper endoscopy, according to the authors' recommendations. This drawback could be corrected by increasing the value of the higher cut-off (i.e., >25). However, this would increase the number of endoscopies required. •Identification of cut-offs is often a critical step in a screening test. The authors selected the thresholds to achieve about 95% NPV and PPV, and thus changed the original cut-offs (18). The robustness of cut-offs (the same or different ones) must be validated extensively by other teams in large cohorts as they might depend on the prevalence of varices in the population. •The P2/MS score may not be reliable in clinical settings such as bacterial infections, which can increase the segmented neutrophil fraction, and in lymphoproliferative diseases, which can increase the lymphocyte fraction. Consequently, the performance of the P2/MS score remains to be established in these situations. •The diagnostic accuracy of the score is quite high. However, this index should be validated by other groups in different parts of the world and in larger series in order to verify this good performance. In addition, the score should be validated in cirrhosis of aetiologies other than HBV (alcohol, HCV, primary biliary cirrhosis, etc.). •It would be interesting to evaluate prospectively whether changes in the score, in the same patients, correspond to changes of the probability of having high-risk varices; should this be the case, follow-up for patients with a low risk of varices could be limited to computing the score at a definite time (e.g. annually). A longitudinal study could address this issue. •Another critical point is the populations to target. Because the prevalence of varices in decompensated cirrhosis is very high (85% in Child C cirrhosis) (1), we suggest that, in these patients, endoscopy should be performed without excluding any patients based on non-invasive predictors. In contrast, in patients with compensated cirrhosis, a noninvasive screening test could be used to rule out patients at a low risk for large varices. Consequently when devising scores to predict varices, the study population should be represented by patients with compensated cirrhosis. On the other hand, it is not always easy to diagnose cirrhosis in compensated patients. In fact, as the authors acknowledge, there is no diagnostic tool that can discriminate between chronic hepatitis and cirrhosis with complete accuracy, the gold standard being the liver biopsy that has a not negligible rate of false negatives (19), and it is not performed in all settings. In fact, development and progression of gastro-oesophageal varices is possible, at least in the HCV setting, even when liver biopsy does not show signs of cirrhosis (20). Therefore, a portion of patients may elude endoscopic screening for varices. This is particularly true in this study where the liver samples could be as low as 15 mm, which is known to lower the accuracy of liver biopsy. Unfortunately, only a few studies have considered all patients with a diagnosis of chronic liver disease (7, 16) – most consider only patients with a diagnosis of cirrhosis. We suggest, in the future, performing the same type of study, but detecting cirrhosis by two different noninvasive methods: one based on noninvasive serum markers or on the index combining laboratory and demographic data, and the second being directly related to the liver, for example transient elastography. In conclusion, the paper by Kim and colleagues provides the validation of a simple score that predicts oesophageal varices in patients with HBV-related cirrhosis with a high accuracy. However, it is necessary to evaluate this score prospectively in this and other settings, to find optimal cut-offs and to compare the diagnostic accuracy of the score with the others. Should the good accuracy in detecting high-risk varices be confirmed and should it be possible to increase the NPV (close to 100%), the P2/MS score could be considered as a surrogate for upper endoscopy in patients with compensated liver disease. In the meantime, we should continue to use, in several situations, what remains, with all its limitations, the gold standard for the detection of gastro-oesophageal varices, namely upper gastrointestinal endoscopy." @default.
• W1923659905 created "2016-06-24" @default.
• W1923659905 creator A5004361418 @default.
• W1923659905 creator A5071429860 @default.
• W1923659905 date "2010-04-08" @default.
• W1923659905 modified "2023-09-25" @default.
• W1923659905 title "Noninvasive prediction of oesophageal varices: as simple as blood count?" @default.
• W1923659905 cites W1966002814 @default.
• W1923659905 cites W1972702648 @default.
• W1923659905 cites W1972912969 @default.
• W1923659905 cites W1980955699 @default.
• W1923659905 cites W1992563722 @default.
• W1923659905 cites W1994356966 @default.
• W1923659905 cites W2006916640 @default.
• W1923659905 cites W2013638983 @default.
• W1923659905 cites W2023824421 @default.
• W1923659905 cites W2080157259 @default.
• W1923659905 cites W2140378669 @default.
• W1923659905 cites W2149981144 @default.
• W1923659905 cites W2157758575 @default.
• W1923659905 cites W2335497705 @default.
• W1923659905 cites W4231481326 @default.
• W1923659905 cites W4254771884 @default.
• W1923659905 doi "https://doi.org/10.1111/j.1478-3231.2010.02317.x" @default.
• W1923659905 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/20707879" @default.
• W1923659905 hasPublicationYear "2010" @default.
• W1923659905 type Work @default.
• W1923659905 sameAs 1923659905 @default.
• W1923659905 citedByCount "4" @default.
• W1923659905 countsByYear W19236599052012 @default.
• W1923659905 countsByYear W19236599052013 @default.
• W1923659905 countsByYear W19236599052014 @default.
• W1923659905 crossrefType "journal-article" @default.
• W1923659905 hasAuthorship W1923659905A5004361418 @default.
• W1923659905 hasAuthorship W1923659905A5071429860 @default.
• W1923659905 hasConcept C111472728 @default.
• W1923659905 hasConcept C126322002 @default.
• W1923659905 hasConcept C138885662 @default.
• W1923659905 hasConcept C2777214474 @default.
• W1923659905 hasConcept C2778012290 @default.
• W1923659905 hasConcept C2780586882 @default.
• W1923659905 hasConcept C71924100 @default.
• W1923659905 hasConcept C90924648 @default.
• W1923659905 hasConceptScore W1923659905C111472728 @default.
• W1923659905 hasConceptScore W1923659905C126322002 @default.
• W1923659905 hasConceptScore W1923659905C138885662 @default.
• W1923659905 hasConceptScore W1923659905C2777214474 @default.
• W1923659905 hasConceptScore W1923659905C2778012290 @default.
• W1923659905 hasConceptScore W1923659905C2780586882 @default.
• W1923659905 hasConceptScore W1923659905C71924100 @default.
• W1923659905 hasConceptScore W1923659905C90924648 @default.
• W1923659905 hasIssue "8" @default.
• W1923659905 hasLocation W19236599051 @default.
• W1923659905 hasLocation W19236599052 @default.
• W1923659905 hasOpenAccess W1923659905 @default.
• W1923659905 hasPrimaryLocation W19236599051 @default.
• W1923659905 hasRelatedWork W146430020 @default.
• W1923659905 hasRelatedWork W2061191495 @default.
• W1923659905 hasRelatedWork W2061829015 @default.
• W1923659905 hasRelatedWork W2080410000 @default.
• W1923659905 hasRelatedWork W2153740656 @default.
• W1923659905 hasRelatedWork W2154362091 @default.
• W1923659905 hasRelatedWork W2416775261 @default.
• W1923659905 hasRelatedWork W2898640880 @default.
• W1923659905 hasRelatedWork W3163747747 @default.
• W1923659905 hasRelatedWork W4289517847 @default.
• W1923659905 hasVolume "30" @default.
• W1923659905 isParatext "false" @default.
• W1923659905 isRetracted "false" @default.
• W1923659905 magId "1923659905" @default.
• W1923659905 workType "article" @default.