Predicting clinical outcomes in patients with bleeding from gastric cancer: novel tools are needed to nail it!

There has been a global trend in the reduction of cancer-related mortality, especially in high-income countries [1]. Primary prevention, screening programs, minimally invasive therapies, targeted chemotherapy, and better palliative treatments can explain, at least in part, this global trend. One of the possible consequences of the longer survival of patients with cancer is an increased frequency of adverse events related to treatment and the disease itself. Gastrointestinal (GI) bleeding is one of the most common causes of admission to the emergency room of oncology centers. In our experience, the bleeding source is the tumor itself in more than 80 % of patients with a primary malignancy located in the upper GI tract [2]. The performance of risk scoring systems in predicting the clinical outcomes of upper GI bleeding from malignancy is unclear. The ideal risk scoring system should accurately identify low risk patients who could be eligible for early discharge, differentiating them from high risk patients who should be managed in the intensive care unit.

“The authors found that the full Rockall score performed better than Glasgow–Blatchford score and admission Rockall score in patients with bleeding from gastric cancer.”

In this issue of Endoscopy, Kim et al. [3] from the National Cancer Center (Gyeonggi, Korea) report on a retrospective observational study in which they compared the performance of three risk scoring systems in predicting the clinical outcomes of patients with GI bleeding from inoperable gastric cancer. From 2001 to 2015, 781 patients with advanced gastric cancer underwent upper GI endoscopy to investigate suspected GI bleeding; 424 patients were excluded (candidates for curative gastrectomy, 62.7 %; incomplete data, 22.9 %; and absence of clinical or endoscopic signs of tumor bleeding, 14.4 %). The Glasgow–Blatchford score (GBS), admission Rockall score, and full Rockall score were calculated for the 357 included patients, who all had inoperable stage IV disease with unequivocal signs of tumor bleeding. Most patients were males in their 50s, receiving palliative chemotherapy, presenting with anemia and melena, and with a large ulcerated tumor involving more than 50 % of the stomach. Adherent clots (Forrest IIb) were the most common endoscopic finding, but high bleeding risk stigmata (i. e. Forrest Ia, Ib, and IIa) were found in roughly one-third of patients. Urgent hemostatic intervention (endoscopic clipping, injection, spraying hemostatic agents, arterial embolization) was performed in 118 patients (33.1 %), and immediate hemostasis was achieved in 108 (93.1 %). There was no mention of the use of hemostatic radiotherapy. The 30-day rebleeding and mortality rates were 17.6 % and 14.8 %, respectively. The main outcome of the study was predicting the need for urgent hemostatic intervention. The authors found that the full Rockall score performed better than the GBS and admission Rockall score. This finding was not surprising considering that the indication for hemostatic intervention was dependent on the endoscopic findings (e. g. gastric cancer with a spurting vessel, Forrest Ia), and only the full Rockall score includes the endoscopic findings in the final score. Restricted to clinical findings, the GBS and admission Rockall score systems performed poorly for predicting the need for hemostatic intervention, with identical AUCs (0.56; 95 %CI 0.50 – 0.62). A GBS of 0 and Rockall score ≤ 6 confidently identified patients who did not require hemostatic intervention. Unfortunately, only four patients (1.1 % of the entire cohort, 1.7 % of the group that did not require hemostatic intervention) had a GBS of 0. Conversely, endoscopy examination is needed for the full Rockall score calculation. Thus, by adopting GBS or admission Rockall score, it is not possible to identify the majority of patients who could be sent home without an endoscopic examination in the first 24 hours.

The clinical implication is clear: endoscopic examination is mandatory for this group of patients. Notably, GBS, admission Rockall score, and full Rockall score performed poorly for the prediction of rebleeding and 30-day mortality, with a minor advantage with the full Rockall score.

The present findings contrast with the high performance of GBS, admission Rockall score, and full Rockall score, especially the former, when GI malignancy is the source of bleeding in less than 5 % of the analyzed cohort [4]. One of the possible interpretations of these discrepant findings is that patients with GI tumor bleeding have a particular behavior, making GBS and the Rockall score less useful than in patients with GI bleeding from benign lesions. Keeping this in mind, we prospectively gauged GBS, Rockall score, and AIMS65 for predicting clinical outcomes in 243 patients with cancer and GI bleeding [5]. AIMS65 is a risk scoring system [6] that evaluates serum albumin, international normalized ratio, mental status, blood systolic pressure, and age > 65 years, and is more accurate than GBS and the admission Rockall score in predicting in-hospital mortality and intensive care unit admission [7]. In our experience, the AIMS65 score performed better in predicting hospital mortality (AUC 0.84) than the other scores [5]. In this sense, it is unfortunate that the authors could not include calculation of the AIMS65 score in their study. Putting these findings into perspective, it is possible that dedicated risk scoring systems need to be developed for predicting clinical outcomes in patients with bleeding from upper GI malignancy. Finally, the impact of novel endoscopic treatments (e. g. hemostatic sprays) on the clinical outcome of patients with bleeding from upper GI tumors remains to be determined [8].

Publication History

Article published online:
22 April 2020

• References

• 1 Torre LA, Siegel RL, Ward EM. et al. Global cancer incidence and mortality rates and trends – an update. Cancer Epidemiol Biomarkers Prev 2016; 25: 16-27
• 2 Maluf-Filho F, Martins BC, de Lima MS. et al. Etiology, endoscopic management and mortality of upper gastrointestinal bleeding in patients with cancer. United European Gastroenterol J 2013; 1: 60-67
• 3 Kim YI, Choi IJ, Lee JY. et al. Comparison of the performance of risk scoring systems for tumor bleeding in patients with inoperable gastric cancer. Endoscopy 2020; 52: 359-367
• 4 Barkun AN, Almadi M, Kuipers EJ. et al. Management of nonvariceal upper gastrointestinal bleeding: guideline recommendations from the International Consensus Group. Ann Intern Med 2019;
• 5 Franco MC, Jang S, Martins BC. et al. Comparison of Glagow-Blatchford, Rockall and AIMS65 scores for predicting upper GI bleeding outcomes in patients with cancer. Gastrointest Endosc 2017; 85: AB69-AB70
• 6 Saltzman JR, Tabak YP, Hyett BH. et al. A simple risk score accurately predicts in-hospital mortality, length of stay, and cost in acute upper GI bleeding. Gastrointest Endosc 2011; 74: 1215-1224
• 7 Robertson M, Majumdar A, Boyapati R. et al. Risk stratification in acute upper GI bleeding: comparison of the AIMS65 score with the Glasgow-Blatchford and Rockall scoring systems. Gastrointest Endosc 2016; 83: 1151-1160
• 8 Chen YI, Wyse J, Lu Y. et al. TC-325 hemostatic powder versus current standard of care in managing malignant GI bleeding: a pilot randomized clinical trial. Gastrointest Endosc 2020; 91: 321-328.e1