Comparison and validation of several scoring systems for non-variceal upper gastrointestinal bleeding: a retrospective study

doi:10.21203/rs.3.rs-4937105/v1

Background

Non-variceal upper gastrointestinal bleeding (NVUGIB) is a serious medical condition that requires effective risk stratification to guide treatment decisions.

Aims

This study aimed to compare and validate the predictive performance of several established scoring systems in patients with NVUGIB: Glasgow–Blatchford score (GBS) and the age, blood tests, and comorbidities (ABC), mental status–anesthesiologist score–pulse–albumin–systolic blood pressure–hemoglobin (MAP(ASH)), Japanese, and Charlson comorbidity index-in-hospital onset–albumin-mental status–Eastern Cooperative Oncology Group performance status–steroids (CHAMPS) scores.

Methods

We retrospectively reviewed the records of 1241 patients with NVUGIB at Chungnam National University Hospital. Each scoring system was evaluated for its ability to predict in-hospital mortality, rebleeding, and the need for radiological or surgical intervention. We also assessed the efficacy of each score in identifying low-risk patients.

Results

The ABC score showed the highest accuracy in predicting in-hospital mortality (C-statistic, 0.890). The MAP(ASH) score was the most effective predictor of rebleeding and the need for interventions (C-statistic, 0.673 and 0.711, respectively). In low-risk patients, the ABC and Japanese scores were the most effective, with very low associated mortality rates.

Conclusions

Different scoring systems have been optimized for various clinical outcomes. The ABC score was the best for predicting mortality, whereas the MAP(ASH) score excelled in identifying rebleeding risks and intervention needs. The selection of an appropriate scoring tool based on specific clinical scenarios can improve patient management and resource allocation in NVUGIB.

Health sciences/Gastroenterology/Gastrointestinal diseases/Gastrointestinal bleeding/Upper gastrointestinal bleeding

Health sciences/Gastroenterology/Gastrointestinal system/Stomach

Upper gastrointestinal bleeding. Endoscopy

Mortality

Scoring systems

Non-variceal upper gastrointestinal bleeding (NVUGIB) is a common and potentially life-threatening medical emergency requiring prompt assessment and management. The mortality rate of NVUGIB ranges from 2–10% depending on patient age, comorbid conditions, and the severity of bleeding.[1, 2] As the population ages and the use of medications such as nonsteroidal anti-inflammatory drugs (NSAIDs) and anticoagulants increases, the incidence and complexity of NVUGIB cases are expected to rise.[3] Clinical guidelines for NVUGIB have established an international consensus for using scoring systems for risk stratification and prognostification.[4–6] To address the challenges of managing NVUGIB, various scoring systems have been developed to assist clinicians in predicting key outcomes such as in-hospital mortality, the need for therapeutic intervention, and the risk of re-bleeding.[7–11] These scoring systems aim to stratify patients based on risk, thereby aiding clinical decision-making and optimizing resource utilization. For instance, the Glasgow–Blatchford Score (GBS) is widely used to predict the need for clinical intervention and identified low-risk patients who can be managed as outpatients.[8, 12, 13] In contrast, the age, blood tests, and comorbidities (ABC) score focuses on predicting mortality with high discriminative power owing to its inclusion of easily measurable parameters.[10, 14] In recent years, new scoring systems such as the Charlson comorbidity index-in-hospital onset–albumin-mental status–Eastern Cooperative Oncology Group performance status–steroids (CHAMPS) and mental status–anesthesiologist score–pulse–albumin–systolic blood pressure–hemoglobin (MAP(ASH)) scores have been introduced, each incorporating unique factors to enhance predictive accuracy. The CHAMPS score includes elements such as the Charlson comorbidity index score and serum albumin levels and can identify patients at low risk of adverse outcomes.[15] The MAP(ASH) score, which includes variables such as mental status and pulse rate, aims to provide a simple yet effective tool for predicting both intervention needs and mortality.[11] Despite the availability of these scoring systems, comprehensive validation studies are required to compare their effectiveness in diverse populations and clinical settings. The strengths and limitations of each scoring system have been highlighted; however, they have not been directly compared in a large cohort of patients with NVUGIB. This study aimed to fill this gap by retrospectively comparing the predictive performance of the ABC score, GBS, MAP(ASH) score, Japanese score, and CHAMPS score in a cohort of patients with NVUGIB treated at a tertiary care center. By evaluating these scoring systems in a real-world clinical setting, we aimed to identify the most reliable and practical tool for predicting the key clinical outcomes of NVUGIB. Ultimately, our goal is to provide clinicians with evidence-based guidance on the optimal use of scoring systems for the management of NVUGIB, thereby improving patient outcomes and healthcare efficiency.

1. Study design and participants

This study was approved by the Institutional Review Board (IRB) of Chungnam National University Hospital (IRB number: CNUH 2024-05-017). The requirement for written informed consent was waived due to the retrospective study design. All procedures performed in studies involving participants were conducted in accordance with the ethical standards of the association, National Research Council, and the 1964 Helsinki Declaration, as amended or equivalent thereto.

This retrospective study was conducted at a tertiary medical center in patients who visited the emergency room (ER) of the Chungnam National University Hospital (CNUH) with suspected upper GI bleeding symptoms. The patients were enrolled between January 2013 and December 2022. NVUGIB was confirmed by endoscopy with one of the following findings: hematemesis, melena, or > 2 g/dL drop in the hemoglobin level compared to that in the prior test. Patients with bleeding from the tumor and those with bleeding after endoscopic resection were excluded. All patients with NVUGIB underwent endoscopic hemostatic therapy, including epinephrine injection, hemoclipping, argon plasma coagulation, and/or hemostatic powder spray. Patients who received hemostasis therapy were hospitalized, and proton pump inhibitor (PPI) therapy was administered immediately. If endoscopic hemostasis failed because of instability at the first visit or during the following endoscopy, vascular embolization or surgery was performed.

2. Data collection and definitions

Data from all patients were collected by the electronic medical record system, and manual chart review was conducted to gather information that could influence the prognosis: patient characteristics (age, sex), initial mental status, initial vital signs (systolic blood pressure, heart rate [HR]), use of medication (NSAIDs, anticoagulants, steroids), ECOG-PS score (Eastern Cooperative Oncology Group performance status), comorbidities (Charlson comorbidity index [CCI]), ASA (American Society of Anesthesiologists) score, and serum test results (hemoglobin, albumin, blood urea nitrogen, creatinine). The scoring systems analyzed were the ABC score, Glasgow–Blatchford score (GBS), MAP(ASH) score, Japanese score, and CHAMPS score. Each scoring system was evaluated using predefined criteria. In-hospital mortality, rebleeding, and radiological or surgical interventions were investigated as outcome data. In-hospital mortality was defined as death that occurred during hospitalization. Rebleeding was defined as recurrent bleeding episodes from the same source with symptoms of vomiting fresh blood, melena, or hemodynamic instability to re-perform endoscopic hemostasis. Radiological or surgical intervention was performed when endoscopic hemostasis failure occurred at the first visit or during hospitalization.

3. Outcomes

The primary outcome was the mortality rate within 30 days of hospital admission due to NVUGIB. This outcome measures the effectiveness of scoring systems in predicting the likelihood of death within a specified period. The secondary outcomes were rebleeding and radiological/surgical intervention rates. The rebleeding rate within in-hospital days was also assessed. Rebleeding was defined as the recurrence of hematemesis, melena, or a significant drop in hemoglobin levels, necessitating additional medical intervention. The need for radiological or surgical intervention was also evaluated. This included any radiological or surgical procedure required to manage ongoing or recurrent bleeding that could not be controlled by endoscopic treatment alone.

4. Statistical analysis

Data were analyzed using IBM SPSS Statistics (version 29.0; IBM Corp., Armonk, NY). Univariate and multivariate logistic regression analyses were performed to determine the possible predictive factors for in-hospital mortality, rebleeding, and radiological/surgical intervention. A P value < 0.05 was determined as statistical significance. The effectiveness of the predictive scoring system was evaluated by calculating the area under the receiver operating characteristic curve (AUROC) using the DeLong test. All the scoring systems were then compared.

1. Patient characteristics

A total of 1241 patients visited the emergency department of CNUH during the study period. Four patients were excluded because of bleeding from the tumor, and 37 patients were excluded because of bleeding after endoscopic resection. (Figure 1)

The baseline characteristics and clinical outcomes of the patients are summarized. (Table 1) Of the 1200 included patients, 901 (75.1%) were men, and the median age of all the patients was 68. Further, the cause of NVUGIB in 1149 (95.7%) patients was peptic ulcer bleeding (65.4%, gastric ulcers; 30.3%, duodenal ulcers). Of all the patients, 156 (13%) experienced rebleeding, 48 (4%) underwent radiological/surgical intervention, 37 (3.2%) died during hospitalization.

Scores including the GBS and ABC, MAP(ASH), Japanese, and CHAMPS scores were calculated for all the included patients, and the median scores are presented. (Table 1)

2. Univariate and multivariate analysis of predictive outcomes

Logistic regression analyses were performed to determine the predictive factors for clinical outcomes such as in-hospital mortality, rebleeding, and radiological/surgical intervention.

Regarding in-hospital mortality, the univariate analysis revealed HR>100bpm, comorbidity of heart failure, ESRD, COPD and liver cirrhosis, initial Hb level, serum albumin level, serum creatinine level, international normalized ratio, ASA score, and ECOG score to be the statistically significant predictive factors, whereas the multivariate analysis suggested initial Hb level, serum albumin level, comorbidities including COPD and liver cirrhosis, ASA, and ECOG score. (Table 2)

Regarding rebleeding, the univariate analysis revealed initial systolic blood pressure, HR, initial Hb level, serum albumin level, prothrombin level, ASA score, and ECOG score to be predictive factors, whereas the multivariate analysis suggested initial systolic blood pressure, HR, serum albumin level, and ECOG score. (Table 3)

Regarding radiological/surgical intervention, the univariate analysis revealed initial systolic blood pressure, HR, serum albumin level, and initial mental status as predictive factors, whereas the multivariate analysis suggested initial systolic blood pressure, heart rate, and serum albumin level. (Table 4)

3. Comparison of the scoring systems

The AUROC was calculated to evaluate the effectiveness of the scoring systems in predicting clinical outcomes. For in-hospital mortality, all scoring systems, except the Japanese scoring system, were effective for prediction. The ABC scoring system was superior to the other scoring systems (C-statistic, 0.890). The C-statistic of the MAP(ASH) score was 0.822 and that of the CHAMPS score was 0.807. The ABC score was superior to the MAP and CHAMPS scores. The discriminative values of the MAP(ASH) and CHAMPS scores were not significantly different. (Figure 2[A]) The most effective scoring system for predicting re-bleeding was the MAP(ASH) score (C-statistic, 0.673) followed by the Japanese and ABC scores (C-statistic = 0.621 for the Japanese score, C-statistic = 0.619 for the ABC scale) (Figure 2[B]). The MAP(ASH) score was the most effective for predicting radiological/surgical intervention (C statistic, 0.711) followed by the ABC and Japanese scores (C statistic, 0.67 and 0.667, respectively). The C-statistic for the CHAMPS score was 0.611. (Figure 2[C])

The number of patients classified as low-risk according to each scoring criterion was as follows: CHAMPS score, 142 (11.8%); GBS, 25 (2.1%); MAP(ASH) score, 315 (26.2%); ABC score, 544 (45.2%); and, Japanese score, 641 (53.2%). The in-hospital mortality rates of these low-risk patients were as follows: CHAMPS score, 0; GBS score, 1 (0.32%); ABC score, 1 (0.18%); and, Japanese score, 17 (2.65%).

We compared and validated several scoring systems for NVUGIB, including the ABC score, Glasgow-Blatchford score (GBS), MAP score, Japanese score, and CHAMPS score. To date, the Glasgow-Blatchford Score (GBS) has shown high discriminative value and potential for clinical use and was particularly helpful in determining the need for intervention and identifying low-risk patients.[8, 12, 13] However, in our study, the GBS demonstrated a lower discriminative value than that in previous studies. One possible reason for this discrepancy is that our study limited the definition of intervention to radiological or surgical, whereas previous studies included a broader range of interventions. Additionally, GBS is difficult to use in clinical practice owing to its complicated calculations. The MAP(ASH) and Japanese scores[16] were also developed to determine the need for intervention, similar to the GBS. The MAP(ASH) score, which includes factors such as mental status, anesthesiology status, and pulse, was designed to predict the need for intervention and mortality regardless of variceal bleeding.[11] It has the advantage of being simple to calculate. The Japanese score predicts the need for therapeutic intervention except for transfusion.[16] Several retrospective studies have validated the Japanese scoring system and compared it with other scoring tools for Korean patients. Choi et al.[17] and Kim et al.[18] reported that the Japanese score is effective for predicting endoscopic intervention but is not superior to other scoring systems for predicting 30-day mortality and rebleeding. However, the calculation complexity remains a limitation. The relatively recently developed ABC score is simple to calculate and has demonstrated high discriminative performance in predicting mortality compared to that of the GBS.[10] Developed in a multicenter prospective study that included various countries and races, the ABC score has shown good performance in several studies.[19, 20] The CHAMPS score (Charlson Comorbidity Index ≥ 2, in-hospital onset, albumin < 2.5 g/dL, altered mental status, Eastern Cooperative Oncology Group performance status ≥ 2, steroids) is a new scoring system developed through a retrospective study in Japan in 2021.[15] It can predict NVUGIB risk stratification and mortality in both outpatient and in-hospital patients before endoscopy. However, the CHAMPS score had a lower performance in our study than that in the original study. This may be due to the exclusion of in-hospital patients from our cohort, in contrast to the original study, which included such patients.

In our analysis, the ABC score exhibited the highest discriminative value for predicting in-hospital mortality. Regarding mortality prediction, Laursen et al.[10] compared the ABC score with AIMS65(age, serum albumin level, systolic blood pressure, prothrombin time (international normalized ratio [INR]), and mental status) for upper gastrointestinal bleeding (UGIB) and found that the ABC score had superior discriminative ability (AUROC 0.81 vs. 0.65). Mules et al.[20] also found that the ABC score demonstrated the highest accuracy for predicting 30-day mortality compared to Progetto Nazionale Emorragia Digestiva(PNED), full Rockall, AIMS65, and GBS in hospitalized patients with UGIB (AUROC 0.85). Similarly, Liu et al.[19] observed that the ABC score outperformed the GBS and AIMS65 in terms of the 90-day mortality rate in patients with UGIB admitted to the emergency room (AUROC 0.72). Marie et al.[21] confirmed the high accuracy of the ABC score for 30-day mortality in patients with UGIB compared to AIMS65 and RS (AUROC, 0.79). While previous studies primarily compared the ABC score with the AIMS65, Rockall, and GBS scores, our study extended the comparison to include the MAP(ASH), Japanese, and CHAMPS scores, showing consistent results. Significant prognostic factors identified through multivariate analysis included initial hemoglobin level, serum albumin level, comorbidities such as COPD and liver cirrhosis, and ASA and ECOG scores. The ABC score incorporates three critical elements: albumin, liver cirrhosis, and the ASA score, which collectively enhance its predictive capability. The MAP(ASH) score also includes the ASA score, albumin level, and hemoglobin level; however, the weight of each factor was greater in the ABC score. The CHAMPS score factors in albumin, ECOG, liver cirrhosis, and COPD. However, because comorbidities are incorporated into the CCI, it is rare for both criteria to be satisfied simultaneously. In addition, the weight assigned to each factor in the CHAMPS score was lower than that assigned in each factor in the ABC score.

In this study, the MAP(ASH) score showed the highest discriminative value for predicting rebleeding and the need for radiological or surgical intervention. Relatively few studies have focused on rebleeding and radiologic/surgical intervention requirements compared to the number of studies investigating mortality outcomes. In particular, because intervention outcomes often include transfusion and endoscopic intervention, our study is unique in comparing the need for radiologic/surgical intervention after the failure of endoscopic intervention. Eduardo et al.[11] found that, regarding rebleeding, the ABC score performed well with GBS in the experimental cohort (AUROC 0.73 vs. 0.72), but it showed lower accuracy than GBS in the validation cohort (0.64 vs. 0.58). Rita et al.[22] reported similar effectiveness of the ABC and MAP(ASH) scores for predicting re-bleeding (AUROC 0.61 vs. 0.67). In our previous study, the ABC, MAP(ASH), and AIMS65 scores showed statistically similar performances for rebleeding and the need for radiologic/surgical interventions.[14] In this study, the prognostic factors for rebleeding identified through the multivariate analysis included the initial systolic blood pressure, heart rate, serum albumin level, and ECOG score. The MAP(ASH) score incorporates three items: pulse, albumin level, and ECOG score. However, the MAP(ASH) score failed to prove clinical usefulness, with an ROC curve C-statistic < 0.7. The prognostic factors for radiological/surgical intervention that emerged through the multivariate analysis were the initial systolic blood pressure, heart rate, and serum albumin level. The MAP(ASH) score including initial low SBP, HR, and serum albumin was the only one that showed an ROC curve C-statistic > 0.7 for radiological/surgical intervention prediction.

The scoring system that included the most significant prognostic factors in each multivariate analysis showed a high predictive power. Unlike in-hospital mortality, scoring systems that showed high discriminative power for rebleeding and radiological/surgical intervention did not show satisfactory results (C-statistic < 0.8).

Although predicting mortality, rebleeding, and the need for intervention are necessary to identify high-risk patients and provide active treatment, appropriate screening of low-risk patients is important to reduce the unnecessary use of medical resources. In this study, the proportion of low-risk patients was 142 (11.8%) with CHAMPS score; 25 (2.1%), GBS score; 315 (26.2%), MAP(ASH) score; 544 (45.2%), ABC score; and, 641 (53.2%), Japanese score. Among them, 0 patients who were assessed based on the CHAMPS score died; 0, GBS score; 1, MAP(ASH) score (0.32%); 1, ABC score (0.18%); and, 17, Japanese score (2.65%). The Japanese score classifies many patients as low risk, but its mortality rate is relatively high compared to those of other scoring systems. The CHAMPS and GBS scores were good at identifying low-risk patients, but the proportion of all patients classified as low risk was relatively low. The ABC and MAP(ASH) scores classified a relatively large number of patients as low-risk and had a very low mortality rate, showing good predictive power.

This study has several strengths. By comparing and analyzing multiple scoring systems, such as the GBS and ABC, GBS, MAP(ASH), Japanese, and CHAMPS scores, we demonstrated the relative strengths and weaknesses of each system. This comprehensive comparison highlights the importance of a tailored approach for the management of patients with NVUGIB. In this study, the ABC and MAP(ASH) scores performed well in predicting prognosis and identifying low-risk patients. The ABC score was particularly useful for predicting 30-day in-hospital mortality in patients with NVUGIB owing to its high prediction accuracy. Conversely, the MAP(ASH) score excelled in predicting rebleeding and the need for surgical intervention. By identifying patients at high risk of rebleeding after the initial bleeding, rebleeding can be minimized through preventive treatment and continuous monitoring. Both scoring systems have the advantage of being easy to calculate, facilitating the identification of low-risk patients during initial treatment. By leveraging the strengths of each scoring system, intensive care can be provided for high-risk patients and resource-efficient management can be implemented for low-risk patients, ultimately improving the prognosis of patients with NVUGIB. Analyzing data of 1241 patients using actual clinical data from a tertiary university hospital increased the practical applicability of our study results. We evaluated the multifaceted utility of each scoring system by analyzing various clinical outcome variables, including 30-day in-hospital mortality, rebleeding, and the need for surgical intervention. Finally, using data from Korean patients helps identify scoring systems that are particularly useful for Asian patients, whereas validating internationally available scoring systems ensures broad applicability.

As this was a retrospective study conducted at a single research institution, there was a risk of selection bias, and the results should be generalized with caution. We attempted to minimize bias by including all patients who met the inclusion criteria for this study. Additionally, the verification was conducted only on outpatients; therefore, it could not be validated whether it could be equally applied to inpatients. Therefore, additional validation studies involving hospitalized patients are required.

Ethics Approval
This retrospective chart review study was approved by the Institutional Review Board (IRB) of Chungnam National University Hospital (IRB number: CNUH 2024-05-017). All procedures performed in studies involving participants were conducted in accordance with the ethical standards of the Association, National Research Council, and the 1964 Helsinki Declaration, as amended or equivalent thereto.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Consent to participate
The requirement for written informed consent was waived due to the retrospective study design.

Funding Sources

No funding was received for the preparation of this manuscript.

Author Contribution

Study concept and design: Hee Seok Moon. Data acquisition: Kyung Ryun In, Hee Seok Moon, and Sun Hyung Kang. Analysis and interpretation of data: Kyung Ryun In and Hee Seok Moon. Drafting of the manuscript: Young Eun Oh and Sun Hyung Kang. Critical revision of the manuscript content: Hee Seok Moon, Jae Kyu Sung, and Hyun Yong Jeong. Statistical analysis: Su Kyoung Jeong and Young Eun Oh.

Data Availability

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding authors.

Laine, L. Upper gastrointestinal bleeding due to a peptic ulcer. N Engl. J. Med. 375, 1198 (2016).
Lee, P. L. et al. Tranexamic acid for gastrointestinal bleeding: A systematic review with meta-analysis of randomized clinical trials. Am. J. Emerg. Med. 45, 269–279 (2021).
Tielleman, T., Bujanda, D. & Cryer, B. Epidemiology and risk factors for upper gastrointestinal bleeding. Gastrointest. Endosc Clin. N Am. 25, 415–428 (2015).
Gralnek, I. M. et al. Endoscopic diagnosis and management of nonvariceal upper gastrointestinal hemorrhage (NVUGIH): European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2021. Endoscopy. 53, 300–332 (2021).
Sung, J. J. et al. Asia-Pacific working group consensus on non-variceal upper gastrointestinal bleeding: an update 2018. Gut. 67, 1757–1768 (2018).
Laine, L., Barkun, A. N., Saltzman, J. R., Martel, M. & Leontiadis, G. I. ACG clinical guideline: upper gastrointestinal and ulcer bleeding. Am. J. Gastroenterol. 116, 899–917 (2021).
Rockall, T. A., Logan, R. F., Devlin, H. B. & Northfield, T. Risk assessment after acute upper gastrointestinal haemorrhage. Gut. 38, 316–321 (1996).
Blatchford, O., Murray, W. R. & Blatchford, M. A risk score to predict need for treatment for upper-gastrointestinal haemorrhage. Lancet. 356, 1318–1321 (2000).
Saltzman, J. R. et al. A simple risk score accurately predicts in-hospital mortality, length of stay, and cost in acute upper GI bleeding. Gastrointest. Endosc. 74, 1215–1224 (2011).
Laursen, S. B. et al. ABC score: a new risk score that accurately predicts mortality in acute upper and lower gastrointestinal bleeding: an international multicentre study. Gut. 70, 707–716 (2021).
Redondo-Cerezo, E. et al. MAP(ASH): A new scoring system for the prediction of intervention and mortality in upper gastrointestinal bleeding. J. Gastroenterol. Hepatol. 35, 82–89 (2020).
Stanley, A. J. et al. Comparison of risk scoring systems for patients presenting with upper gastrointestinal bleeding: international multicentre prospective study. BMJ. 356, i6432 (2017).
Orpen-Palmer, J. & Stanley, A. J. Update on the management of upper gastrointestinal bleeding. BMJ Med. 1, e000202 (2022).
Sakong, H. et al. ABC score is an effective predictor of outcomes in peptic ulcer bleeding. Med. (Baltim). 101, e31541 (2022).
Matsuhashi, T. et al. A simple prediction score for in-hospital mortality in patients with nonvariceal upper gastrointestinal bleeding. J. Gastroenterol. 56, 758–768 (2021).
Iino, C. et al. Evaluation of scoring models for identifying the need for therapeutic intervention of upper gastrointestinal bleeding: A new prediction score model for Japanese patients. Dig. Endosc. 28, 714–721 (2016).
Matsuhashi, T. et al. Validation and comparison of the Japanese score and other scores in patients with nonvariceal upper gastrointestinal bleeding: a retrospective study. Dig. dis. 41, 66–73 (2023).
Kim, S. H. et al. Comparison and validation of the Japanese score and other scoring systems in patients with peptic ulcer bleeding: A retrospective study. Med. (Baltim). 102, e34986 (2023).
Liu, S., Zhang, X., Walline, J. H., Yu, X. & Zhu, H. Comparing the performance of the ABC, AIMS65, GBS, and pRS scores in predicting 90-day mortality or rebleeding among emergency department patients with acute upper gastrointestinal bleeding: a prospective multicenter study. J. Transl Int. Med. 9, 114–122 (2021).
Mules, T. C. et al. Comparison of risk scoring systems in hospitalised patients who develop upper gastrointestinal bleeding. GastroHep. 3, 5–11 (2021).
Saade, M. C. et al. Validation of the new ABC score for predicting 30-day mortality in gastrointestinal bleeding. BMC Gastroenterol. 22, 301 (2022).
Jimenez-Rosales, R. et al. Performance of the new ABC and MAP(ASH) scores in the prediction of relevant outcomes in upper gastrointestinal bleeding. J. Clin. Med. 12, 1085 (2023).

Table 1. Baseline characteristics of the study population

Characteristic	No. (%) (n=1200)
Male sex	901 (75.1)
Age, mean±SD, y	66.6±27.7
≥65	696 (56.8)
≤64	504 (41.1)
Cause of NVUGIB
Gastric ulcer	785(65.4)
Duodenal ulcer	364(30.3)
Other	51(4.3)
Initial vital sign
SBP, mean±SD, mmHg	116±23.5
DBP, mean±SD, mmHg	66.8±13.7
Heart rate, mean±SD, bpm	92.4±20.3
Comorbidity
Hypertension	610(50.8)
Diabetes mellitus	334(27.8)
Myocardial infarction	119(9.9)
Heart failure	89(7.4)
ESRD	92(7.7)
COPD	32(1.8)
Liver cirrhosis	77(6.4)
Blood tests
Hemoglobin, mean±SD, g/dL	8.9±2.6
Albumin, mean±SD, g/dL	3.2±0.6
BUN, mean±SD ,mg/dL	43.7±26.5
Creatinine, mean±SD, mg/dL	1.5±2.0
PT(INR), mean±SD	1.2±0.6
Medication
Antiplatelet agents	456(38.0)
NSAIDs	59(4.9)
Steroid	17(1.4)
Re-bleeding	156(13.0)
Radiological intervention	48(4.0)
In-hospital mortality	37(3.2)
Scoring system
MAP score, mean±SD	2.7±1.7
ABC score, mean±SD	4.3±2.8
Japanese score, mean±SD	2.4±2.3
GBS, mean±SD	10.3±3.4
CHAMPS score, mean±SD	1.4±0.9

NVUIB, non-variceal upper gastrointestinal bleeding; SBP, systolic blood pressure; DBP, diastolic blood pressure; ESRD, End stage renal disease; COPD, chronic obstructive pulmonary disease; BUN, blood urea nitrogen; PT, prothrombin time; INR, international normalized ratio; NSAIDs, nonsteroidal anti-inflammatory drugs; MAP, mental status–anesthesiologist score–pulse; ABC, age-blood tests-comorbidity; GBS, Glasgow–Blatchford score; CHAMPS, Charlson comorbidity index-in-hospital onset–albumin–mental status–Eastern Cooperative Oncology Group performance status–Steroids

Table 2. Logistic regression analysis of predictive factors for in-hospital mortality

Characteristic	In-hospital mortality (N=57)	p-value	Multivariate analysis
			OR(95% CI)			p-value
Female sex	18 (31.6)	0.236
Age, y		0.105
≤64	18 (31.6)
≥65	39 (61.4)
SBP		0.111
≥100 mmHg	38 (66.7)
<100 mmHg	19 (33.3)
HR		0.041
≤100 bpm	30 (52.6)
＞100 bpm	27 (47.4)
Comorbidity
HTN	24 (42.1)	0.179
DM	17 (29.8)	0.731
MI	9 (15.8)	0.133
CHF	10 (17.5)	0.004
ESRD	11 (19.3)	0.001
COPD	6 (10.5)	<0.001		4.082(1.408-11.84)	0.01
LC	13 (22.8)	<0.001		2.583(1.159-5.757)	0.02
Medication
Antiplatelet	20 (35.1)	0.643
Steroid	1 (1.8)	0.859
Hemoglobin		<0.001		5.014(1.503-16.728)	0.009
≥10 mg/dL	3 (5.3)
<10 mg/dL	54 (94.7)
Albumin		<0.001		2.453(1.261-4.772)	0.008
≥2.5 g/dL	34 (59.6)
<2.5 g/dL	23 (40.4)
BUN		0.418
<22.4 mg/dL	8 (14.0)
≥22.4 mg/dL	49 (86.0)
Creatinine		<0.001
<1.7 mg/dL	37 (64.9)
≥1.7 mg/dL	20 (35.1)
INR		<0.001
<1.5	42 (73.7)
≥1.5	15 (26.3)
Altered mentality	7 (12.3)	0.07
ASA score		<0.001		3.071(1.203-7.844)	0.019
<3	6 (10.5)
≥3	51 (89.5)
ECOG-PS score		<0.001		3.375(1.817-6.269)	<0.001
<2	30 (52.6)
≥2	27 (47.4)

OR, odds ratio; CI, confidence interval; SBP, systolic blood pressure; ESRD, end-stage renal disease; COPD, Chronic obstructive pulmonary disease; BUN, blood urea nitrogen; PT, prothrombin time; INR, international normalized ratio; NSAIDs, nonsteroidal anti-inflammatory drugs; ASA, American Society of Anesthesiologists; ECOG-PS, Eastern Cooperative Oncology Group performance status.

Table 3. Logistic regression analysis of predictive factors for rebleeding

Characteristic	Re-bleeding (N=156)	p-value	Multivariate analysis
			OR(95% CI)			p-value
Female sex	31 (19.9)	0.12
Age, y		0.139
≤64	57 (36.5)
≥65	99 (63.5)
SBP		<0.001		1.949(1.341-2.832)	<0.001
≥100 mmHg	93 (59.6)
<100 mmHg	63 (40.4)
HR		0.004		1.565(1.089-2.249)	0.016
≤100 bpm	86 (55.1)
＞100 bpm	70 (44.9)
Comorbidity
HTN	72 (46.2)	0.211
DM	48 (30.8)	0.381
MI	18 (11.5)	0.468
CHF	16 (10.3)	0.149
ESRD	18 (11.5)	0.054
COPD	6 (3.8)	0.331
LC	12 (7.7)	0.487
Medication
Antiplatelet	59 (37.8)	0.961
Steroid	4 (2.6)	0.653
Hemoglobin		0.006
≥10 mg/dL	35 (22.4)
<10 mg/dL	121 (77.6)
Albumin		<0.001		2.093(1.313-3.335)	0.002
≥2.5 g/dL	116 (74.4)
<2.5 g/dL	40 (25.6)
BUN		0.787
<22.4 mg/dL	27 (17.3)
≥22.4 mg/dL	129 (82.7)
Creatinine		0.264
<1.7 mg/dL	125 (80.1)
≥1.7 mg/dL	31 (19.9)
INR		0.013
<1.5	135 (86.5)
≥1.5	21 (13.5)
Altered mentality	12 (7.7)	0.487
ASA score		<0.001
<3	51 (32.7)
≥3	105 (67.3)
ECOG-PS score		<0.001		2.262(1.464-3.496)	<0.001
<2	112 (71.8)
≥2	44 (28.2)

OR, odds ratio; CI, confidence interval; SBP, systolic blood pressure; ESRD, end-stage renal disease; COPD, chronic obstructive pulmonary disease; BUN, blood urea nitrogen; PT, prothrombin time; INR, international normalized ratio; NSAIDs, nonsteroidal anti-inflammatory drugs; ASA, American Society of Anesthesiologists; ECOG-PS, Eastern Cooperative Oncology Group performance status; HTN, hypertension; MI, myocardial infarction; DM, diabetes mellitus; CHF, congestive heart failure; LC, liver cirrhosis; BUN, blood urea nitrogen; HR, heart rate

Table 4. Logistic regression analysis of predictive factors for radiological/surgical intervention

Characteristic	Re-bleeding (N=48)	p-value	Multivariate analysis
			OR(95% CI)			p-value
Female sex	9 (18.8)	0.315
Age, y		0.069
≤64	14 (29.2)
≥65	34 (70.8)
SBP		<0.001		1.949(1.341-2.832)	<0.001
≥100 mmHg	19 (39.6)
<100 mmHg	29 (60.4)
HR		<0.001		1.565(1.089-2.249)	0.016
≤100 bpm	20 (41.7)
＞100 bpm	28 (58.3)
Comorbidity
HTN	28 (58.3)	0.288
DM	14 (29.2)	0.836
MI	5 (10.4)	0.907
CHF	5 (10.4)	0.422
ESRD	7 (14.6)	0.068
COPD	3 (6.3)	0.130
LC	5 (10.4)	0.255
Medication
Antiplatelet	14 (29.2)	0.2
Steroid	1 (2.1)	0.999
Hemoglobin		0.284
≥10 mg/dL	12 (25.0)
<10 mg/dL	36 (75.0)
Albumin		<0.001		2.093(1.313-3.335)	0.002
≥2.5 g/dL	30 (62.5)
<2.5 g/dL	18 (37.5)
BUN		0.308
<22.4 mg/dL	6 (12.5)
≥22.4 mg/dL	42 (87.5)
Creatinine		0.695
<1.7 mg/dL	39 (81.3)
≥1.7 mg/dL	9 (18.8)
INR		0.281
<1.5	42 (87.5)
≥1.5	6 (12.5)
Altered mentality	7 (14.6)	0.023
ASA score		0.141
<3	17 (35.4)
≥3	31 (64.6)
ECOG-PS score		0.056
<2	37 (77.1)
≥2	11 (22.9)

OR, odds ratio; CI, confidence interval; SBP, systolic blood pressure; ESRD, end-stage renal disease; COPD, chronic obstructive pulmonary disease; BUN, blood urea nitrogen; PT, prothrombin time; INR, international normalized ratio; NSAIDs, nonsteroidal anti-inflammatory drugs; ASA, American Society of Anesthesiologists; ECOG-PS, Eastern Cooperative Oncology Group performance status; HTN, hypertension; MI, myocardial infarction; DM, diabetes mellitus; CHF, congestive heart failure; LC, liver cirrhosis ; BUN, blood urea nitrogen; HR, heart rate

No competing interests reported.

Comparison and validation of several scoring systems for non-variceal upper gastrointestinal bleeding: a retrospective study

Status:

Version 1

Abstract

Background

Aims

Methods

Results

Conclusions

Figures

Introduction

Materials and Methods

1. Study design and participants

2. Data collection and definitions

3. Outcomes

4. Statistical analysis

Results

Discussion / Conclusion

Declarations

Conflict of Interest Statement

Consent to participate
The requirement for written informed consent was waived due to the retrospective study design.

Funding Sources

Author Contribution

Data Availability

References

Tables

Additional Declarations

Status:

Version 1

Comparison and validation of several scoring systems for non-variceal upper gastrointestinal bleeding: a retrospective study

Status:

Version 1

Abstract

Background

Aims

Methods

Results

Conclusions

Figures

Introduction

Materials and Methods

1. Study design and participants

2. Data collection and definitions

3. Outcomes

4. Statistical analysis

Results

Discussion / Conclusion

Declarations

Conflict of Interest Statement

Consent to participate The requirement for written informed consent was waived due to the retrospective study design.

Funding Sources

Author Contribution

Data Availability

References

Tables

Additional Declarations

Status:

Version 1

Consent to participate
The requirement for written informed consent was waived due to the retrospective study design.