a b s t r a c t

Introduction: Upper gastrointestinal bleeding is one of the common causes of mortality and morbidity. The Rock- all score (RS) and Glasgow-Blatchford score (GBS) are frequently used in determining the prognosis and predicting in-hospital adverse events, such as mortality, re-bleeding, hospital stay, and blood transfusion require- ments. The quick Sepsis Related Organ Failure Assessment (qSOFA) score is easy and swift to calculate. The com- monly used scores and the qSOFA score were compared and why and when these scores are most useful was investigated.

Method: 133 patients admitted to the emergency department with upper gastrointestinal bleeding over the pe- riod of a year, were evaluated in this retrospective study. The RS, GBS and qSOFA score were calculated for each patient, and their relationship with in-hospital adverse events, such as length of hospitalization, rebleeding, en- doscopic treatment, blood Transfusion requirements, and mortality, was investigated.

Results: The mean overall GBS was 9.72 +- 3.72 (0-19), while that of patients who did not survive was 14.0 +- 1.1 (13-16), with an area under the curve (AUC) of 0.901, a cutoff value of 12.5, and specificity (Spe) and sensitivity (Sen) of 1 and 0.82, respectively. The median value of the GBS, in terms of transfusion need, was 7.12 +- 4.01 (0-15).

(AUC = 0.752, cut-off = 9.5, Spe = 0.79, Sen = 0.69). The median value of the qSOFA score, in terms of intensive care need, was 1.73 +- 0.7 (0-3) (AUC = 0.921, cut-off = 0.5, Spe = 0.93, Sen = 0.79). The RS median, in terms of re-bleeding, was 8.22 +- 0.97 (6-9).

Conclusion: early use of Risk stratification scores in upper gastrointestinal bleeding is important due to the high risk of morbidity and mortality. All scoring systems were effective in predicting mortality, the need for intensive care, and re-bleeding. The GBS had a greater Predictive power in terms of mortality and transfusion need, the qSOFA score for intensive care need, and the RS for re-bleeding. The simpler, more efficient, and more easily cal- culated qSOFA score can be used to estimate the severity of patients with upper gastrointestinal bleeding.

(C) 2021

1. Introduction

Upper gastrointestinal (GI) bleeding is one of the primary causes of emergency department admissions, and it includes bleeding into the in- testinal lumen from anywhere between the upper esophagus and the Treitz ligament [1]. Its annual incidence is 50-172 per 100,000, with high mortality, morbidity, and Hospitalization costs. Even if the fre- quency of upper GIS bleeding decreases as a result of preventive mea- sures, mortality remains unchanged. Despite improvements in

* Corresponding author.

E-mail address: [email protected] (B. Taslidere).

diagnosis and treatment, its mortality is between 2 and 8% according to the ABC score. ABC score (age, blood test, comorbidity) is the latest score and best easy score for predicting mortality [2,3]. Upper GIS bleed- ing presents with a wide spectrum of symptoms, ranging from an occult hemorrhage to heavy bleeding, and from dizziness to hypovolemic shock [4]. Rapid diagnosis and accurate classification are crucial as it is a life-threatening disease [5]. Various scoring systems are used to pre- dict in-hospital adverse events, such as mortality, re-bleeding, length of hospital stay, intensive care, and the need for blood transfusion. In- hospital adverse events not only cause direct harm to the patient but also constitute a Financial burden on the healthcare system. Therefore, a thorough examination of such patients is required. Scoring systems

https://doi.org/10.1016/j.ajem.2021.02.016

0735-6757/(C) 2021

form part of an integral approach to GIS bleeding. The most frequently used scoring methods include the Glasgow-Blatchford score and Rockall Score (RS) [6]. The RS–developed to evaluate the risk of re- bleeding and mortality–evaluates age, shock findings, accompanying diseases, endoscopic diagnosis, and the cause of the last bleed. This sys- tem helps identify low-risk patients and their early discharge. Three pa- rameters are used to calculate the pre-endoscopic Rockall score (PERS): Age, shock findings, and the presence of comorbidities. It can be used to reduce the need for urgent endoscopy. The GBS was developed to pre- dict the risk of complications in patients [7,8]. Clinical and laboratory findings are used to calculate this score, and endoscopic findings are not taken into consideration–the presence of blood urea, hemoglobin, blood pressure, and comorbidities determine the score. Many clinical and laboratory parameters are required to calculate the RS and GBS, and these take time, rendering them difficult to use in emergency ser- vices. In some cases, bleeding stops spontaneously; however, in some, it continues or is recurring [7]. Emergency physicians responsible for acute Patient evaluation may find it difficult to discharge such patients without endoscopy; however, it is risky for both patients and physicians when patients with life-threatening conditions are required to wait in the emergency department for long periods of time. The search for a sat- isfactory, fast, and reliable score that can be used in upper GIS bleeding and can be diagnosed by anamnesis and physical examination, has prompted this study. The quick Sepsis Related Organ Failure Assessment (qSOFA) score, consisting of parameters already examined at the triage stage, is easy to remember and calculate for all patients admitted to the emergency department. It may also be more favorable than the other scoring methods, in patients whose comorbidities are unknown and shock findings are not evaluated. The only criteria considered are respi- ratory rate, altered state of consciousness, and systolic blood pressure

[9] (Table 1).

The qSOFA score, which promptly evaluates the severity of the dis- ease, can reduce the time, cost, and length of hospital stay by predicting in-hospital adverse events in patients with acute upper gastrointestinal (GI) bleeding. This is achieved by ensuring the correct use of resources in emergency situations. This study comparatively investigated why, when, and in which situations RS, PERS, GBS, and qSOFA scores could be used.

1. Method

This study was conducted retrospectively, in patients over 18 years of age, who were admitted to the emergency department between 01 September 2018 and 31 August 2019. Information concerning the pa- tients was obtained from emergency service forms and the hospital reg- istry. The codes associated with upper GIS bleeding were examined in the hospital database. Patients with varicose bleeding, those who did

not undergo endoscopy or whose bleeding was unrelated to upper GIS, trauma patients, pregnant women, and patients with incomplete data were excluded from the study. Clinical and laboratory data were collected retrospectively, including endoscopic findings, laboratory findings, treatment and clinical follow-up. Thus, 133 out of 246 patients were included. Fig. 1 shows the patients included and excluded from the study. Patients’ age, gender, chronic diseases, Presenting complaints, medications, symptoms (hematochezia, hematemesis, melena, syn- cope, active bleeding), blood pressure, rectal examination findings, kid- ney function tests, coagulation tests, prothrombin time, activated thromboplastin time, international normalized ratio values, hemoglobin, hematocrit, thrombocyte values, need and number of blood transfusions, endoscopy findings, endoscopic interventions, inci- dence of re-bleeding, duration of hospital stay, and outcomes (dis- charge, service/intensive care admission, death) were recorded on pre-created forms. The PERS and endoscopic RS, GBS, and qSOFA score were calculated for each patient; how these scores changed in non- survivors and patients who developed in-hospital adverse events were investigated. The qSOFA score range was 0-3, PERS 0-7, RS 0-11, and GBS 0-29. In the follow-up of the patients, hematemesis, melena, and presence of hemodynamic instability criteria indicated recurrent bleed- ing. In-hospital adverse events included the need for transfusions, re- bleeding, mortality, intensive care, and hospitalizations longer than five days. The accuracy of the scoring systems was assessed by plotting receiver-operating characteristic curves (ROC curves). As the number of patients who died as a result of upper GI bleeding either decreased or remained static with each passing year, all data meeting the relevant conditions for the focus year, were scanned retrospectively, and in- cluded in the study.

The approval of the Ethics Committee of the University (21/01/2020,

issue 02/33) was obtained to conduct the study.

1. Outcomes

Outcomes were retrospectively assessed by reviewing of the hospi- tal medical database. The primary study outcomes: Using a score that can predict the intensive care needs of patients with upper GI bleeding. Secondary outcomes: To determine what the most useful scores are in terms of “in-hospital adverse events” (mortality, recurrent bleeding and need for transfusion, etc.).

1. Statistical analysis

Behaviors of quantitative variables were expressed using centraliza- tion and variance measures, with Mean +- SD. To show the behavioral differences of group mean values, ANOVA t-test was used in cases where normality and uniformity assumptions were met, and Mann-

Table 1

Characteristics of scoring systems.

Scoring systems

Year of introduction

Parameters Characteristics

GBS 2000 Hemoglobin This system has use in prediction of outcomes such as probability of re-bleeding and need for interventions

Systolic blood pressure Pulse

Blood urea nitrogen Melena or syncope

Liver disease or heart failure

RS 1996 Age, co-morbidity, shock findings,

endoscopic findings

like endoscopy, surgery, and blood transfusion

This system is valuable to predict re-bleeding and mortality rates for patients with upper gastrointestinal bleeding

PERS 1999 Age, co-morbidity, shock findings Stratifying patients to determine low risk patients can allow discharge from the emergency department with

planned outpatient endoscopy

qSOFA 2016 Respiratory rate This system is used at the bedside to help identify and risk stratify patients with sepsis altered mentation

Systolic blood pressure

GBS: Glasgow-Blatchford Score; RS: Rockall Score; PERS: Pre-endoskopik Rockall Score; qSOFA: Quick Sequential Organ Failure Assesment.

Fig. 1. Patients inclusion in the study.

Whitney U Test (number of groups = 2) non-parametric method was used otherwise. ROC (receiver Operator characteristics) curve analysis was used to find the cut-off values of the categorical variables based on numerical values. P = 0.05 was considered statistically significant in all cases. Statistical analysis was performed with the IBM SPSS (Statis- tics Package for Social Sciences for Windows, Version 21.0, Armonk, NY, IBM Corp.) package program.

1. Missing data

The prevalence and pattern of missing data was evaluated and found to be missing completely at random (Little’s test: p = 0,0.085 > 0.05). Missing data in the main cohort were handled by exluding these pa- tients which has 39% of overall.

1. Sample size determination

The study had between 50.9%-100% power to produce a significant difference with N = 133 participants in terms of GBS, RS, PERS, Quick SOFA with 0.05 type 1 error, and the mean of power is found as 93,1 +- 17,1 (Table 2).

1. Results

Of the 133 patients included in the study, 62.9% were male (n = 83) and 37.1% were female (n = 49), with an overall mean age of 63.1 +-

17.29 (20-91) years. The mean age of Males and females were 59.7 +-

17.8 years and 69 +- 14.6 years, respectively (p = 0.02) (Table 3). Pa- tients’ complaints at presentation included dyspepsia and heartburn (n = 51, 38.6%), abdominal pain (n = 37, 28%), nausea/vomiting

(n = 32, 24.2%), dizziness (n = 6, 4.5%), and syncope (n = 6, 4.5%). The types of bleeding were melena (n = 95, 72%), hematemesis (n = 20, 15.2%), hematochezia (n = 11, 8.3%), and active bleeding (n = 6, 4.5%). Of these patients, 38 (28.8%) were not using any medication, while 37 were using Antiplatelet agents (28%), 20 were using anticoag- ulants (15.2%), six were using new generation anticoagulants (4.5%), six were using Non-steroidal anti-inflammatory drugs (NSAIDs) (4.5%), and 25 patients were using other drugs (18.9%). Endoscopic epinephrine

Table 2

Sample size determination.

Power	GBS	RS	PERS	Quick SOFA
Mortality	100%	50.9%	98.7%	95.6%
ICU	100%	99.8%	99.8%	100%

GBS: Glasgow-Blatchford Score; RS: Rockall Score; PERS: Pre-endoskopik Rockall Score; qSOFA: Quick Sequential Organ Failure Assesment, ICU: Intensive care unit.

injection was performed on 99 of the patients (75%), twelve patients re- ceived combined therapy (9%) and argon plasma coagulation was per- formed on a further seventeen patients (12.7%1) and five patients received other therapy (endoscopic clipping or electrocautery) (3.7%). Endoscopy findings included a Duodenal ulcer (n = 40, 30.3%), gastric ulcer (n = 32, 24.2%), no lesions (n = 19, 14.4%), malignancy (n = 10, 7.6%), angiodysplasia (n = 9, 6.8%), erosive hyperemic gastritis

(n = 7, 5.3%), ulcerated esophagitis (n = 6, 4.5%), none identified

(n = 6, 4.5%), and Mallory-Weiss (n = 3, 2.3%).

Of the patients investigated, 11.4% (n = 15) were discharged, while 102 were hospitalized (77.3%), nine were admitted to the intensive care unit (6.8%), and six died (4.5%). Comorbidities were present in 91 (68.9%) patients, including hypertension (n = 58, 43.9%), diabetes mellitus (n = 33, 25%), coronary artery disease (n = 33, 25%), heart fail- ure (n = 15, 11.4%), kidney disease (n = 13, 9.8%), liver disease (n = 11, 8.3%), malignancy (n = 8, 6.1%), and Cerebrovascular events (n = 4, 3%). Thirty-five (26.5%) patients had a history of previous bleeding. Rectal examination findings were present in 113 (85.6%) patients and 126 (95.5%) patients were conscious with The Glasgow Coma Scale of 15 points. Transfusion was required in 33 (25%) patients and re-bleeding was observed in nine patients (6.8%). The mean number of transfusions was 2.1 +- 2 (0-12). In-hospital mortality was observed in six (4.5%) patients.

The mean GBS was 9.7 +- 3.7 (0-19), mean RS was 4.1 +- 2.1, (0-9),

mean PERS was 2.6 +- 1.6, (0-6), and mean qSOFA score was 0.4 +- 0.7,

Table 3

Clinical characteristics and laboratory values of the study patients.

	Units	Mean +- SD	Median (IQR25–75)
Age	Years	63.16 +- 17.29	66 (20-91)
Sex, number/age	Female	49 (37.1%)/69.02 +- 14.647	69 (20-91)
	Male	83 (62.9%)/59.70 +- 17.877	62 (22-88)
WBC	10^3/uL	8.98 +- 3.87	8.8 (2.5-18)
Hemoglobin	g/Dl	9.02 +- 1.85	9 (3.84-15.2)
Hematocrit	%	27.96 +- 5.39	27.9 (11.7-45.6)
MCV	fL	85.81 +- 7.72	87 (30-100)
Albumin	g/dL	3.19 +- 0.64	3.2 (1.7-4.7)
BUN	mg/dL	30.81 +- 25.75	27 (5-175)
Creatinine	mg/dL	1.08 +- 1.19	0.82 (0.13-12)
BCR		30.81 +- 25.75	27 (5-175)
Prothrombin time	Sec.	15.8 +- 2.37	15.2 (12-24)
Platelet	10^3/uL	263.03 +- 106.38	249 (30-758)
INR		2.02 +- 2.83	1.23 (0.88-22.49)
Sodium	mmol/L	137.16 +- 2.74	137 (128-144)
Poatassium	mmol/L	4.17 +- 0.55	4.19 (2.7-6.03)
LDH	U/L	201.2 +- 70.01	183 (92-537)

WBC: White Blood Cell, LDH: Lactate dehydrogenase, BCR: Blood urea nitrogen/creatinine ratio, INR: International Normalized Ratio, MCV: Mean corpuscular volume, BUN: Blood urea nitrogen.

(0-3). In the patients who died, the mean value of GBS was 14 +- 1.1 [13-16], the mean RS value was 6 +- 2.5 [2-9], the mean PERS was

4.6 +- 1.2 [3-6], and the qSOFA score was 1.8 +- 0.9 (0-3). In the surviv- ing patients, the mean value of GBS was 10 (0-19) (p < 0.001), the mean RS value was 4 (0-9) (p = 0.05), the mean PERS was 2 (0-6) (p = 0.04), and the qSOFA score was 0 (0-2) (p < 0.001) (Table 4).

For those patients requiring intensive care, the median GBS was

13.5 +- 2.3 [9-19], the median RS was 6.4 +- 1.9 [2-9], the median

PERS was 4.2 +- 1.3 [2-6], and the median qSOFA score was 1.7 +- 0.7 (0-3). In those patients not requiring intensive care, the median GBS was 10 (0-17) (p < 0.001), the median RS was 4 (0-9) (p < 0.001), the median PERS was 2 (0-6) (p < 0.001), and the median qSOFA score was 0 (0-2) (p < 0.001). Intensive care: GBS (AUC = 0.856, cut- off >= 11.5, Spe = 0.87, Sen = 0.74), RS (AUC = 0.828, cut-off >= 4.5, Spe = 0.87, Sen = 0.68), PERS (AUC = 0.796, cut-off >= 3.5, Spe = 0.73, Sen = 0.73), and qSOFA (AUC = 0.921, cut-off >= 0.5, Spe = 0.93,

Sen = 0.79), (Table 5), (Figs. 2, 3).

The median value, AUC, cut-off, the specificity and the sensitivity values of the scores in terms of need for mortality, transfusion, re- bleeding and in-hospital adverse events are shown in figures (Figs. 2, 4-6).

1. Discussion

Upper GI bleeding can be a life-threatening condition and requires careful evaluation in order to reduce the risk of rebleeding, hemorrhagic shock, and death. The outcome of these patients depends on the resus- citation measures taken during admission to the hospital and an ade- quate assessment of their risk levels. For this purpose, the best-known scores among those used so far are the GBS and the RS [10,11]. These widely used scores were developed to determine the need for clinical intervention, transfusion, and endoscopy, along with the likelihood of mortality and re-bleeding [12]. Although the GBS and the RS can deter- mine the need for clinical intervention in patients, they are not suffi- ciently useful for predicting intensive care needs. This is why Budimir et al. [7] concluded that there is no such thing as a perfect score and that the best option is to use multiple scoring systems at the same time [13,14]. The main objective is to determine the need for intensive care in patients with upper GI bleeding in the early stages. In this study, the rate of patients requiring intensive care was found to be 6.8%, although some existing studies suggest that this rate can be as high as 20% [13]. As acute upper GI bleeding is treated more effectively with medical and endoscopic methods, there is, however, a decrease in mortality and the need for intensive care [14,15].

Some studies have found that those with a GBS of 13 and above are highly likely to be hospitalized in intensive care, whereas other studies suggest those with a score of 12 and above [16,17]. In this current study, the GBS cut-off value for patients needing intensive care was 11.5, with a median value of 13.53 +- 2.33. As can be seen, the GBS must exceed the high cut-off value to predict the need for intensive care. However, an ideal scoring system should be simpler, more practical, and easier to

Table 4

Vital signs and score averages.

	Units	Mean +- SD	Median (IQR25–75)
Systolic Blood Pressure	mmHg	106.55 +- 14.17	110 (60-150)
Diastolic blood pressure	mmHg	61.73 +- 12.15	60 (30-90)
Heart rate	bpm	93.8 +- 18.96	90 (58-143)
Respiratory rate	/min	19.17 +- 1.35	20 (16-24)
Blood transfusion	number	2.16 +- 2.02	2 (0-12)
GBS	Score	9.72 +- 3.72	10 (0-19)
RS	Score	4.16 +- 2.11	4 (0-9)
PERS	Score	2.62 +- 1.65	3 (0-6)
Quick SOFA	Score	0.42 +- 0.72	0 (0-3)

GBS: Glasgow-Blatchford Score, RS: Rockall score, PERS: Pre-endoscopy Rockall score.

Table 5

ROC values of all risk scores for prediction of clinical outcomes.

	GBS	RS	PERS	Quick SOFA
Mortality Intensive care	Auc = 0.901 Spe = 1 Sen = 0.82 Auc = 0.856	Auc = 0.735 Spe = 0.83 Sen = 0.63 Auc = 0.828	Auc = 0.844 Spe = 0.83 Sen = 0.7 Auc = 0.796	Auc = 0.866 Spe = 0.83 Sen = 0.91 Auc = 0.921
Re-bleeding Blood transfusion	Spe = 0.87 Sen = 0.74 Auc = 0.759 Spe = 0.78 Sen = 0.71 Auc = 0.752 Spe = 0.79 Sen = 0.69	Spe = 0.87 Sen = 0.68 Auc = 0.981 Spe = 0.89 Sen = 0.99 Auc = 0.4 Spe = 0 Sen = 1	Spe = 0.73 Sen = 0.73 Auc = 0.885 Spe = 0.78 Sen = 0.89 Auc = 0.65 Spe = 0.52 Sen = 0.74	Spe = 0.93 Sen = 0.79 Auc = 0.848 Spe = 0.89 Sen = 0.76 Auc = 0.481 Spe = 0.15 Sen = 0.89
Adverse events	Auc = 0.822 Spe = 0.69 Sen = 0.89	Auc = 0.72 Spe = 0.45 Sen = 0.86	Auc = 0.741 Spe = 0.76 Sen = 0.64	Auc = 0.597 Spe = 0.33 Sen = 0.86

use. Both the GBS and the RS are currently in wide use, yet they are com- plex and time-consuming as they require numerous parameters, some of which rely on lab results [18]. Moreover, the RS requires endoscopic data for computation, which is impossible to perform at presentation. The qSOFA score, which has not been used in patients with upper GI bleeding until now, is a simple score consisting of just three parame- ters–respiratory rate, mental state, and systolic blood pressure [19]. Many studies have shown that the qSOFA score can be used to predict the need for intensive care and likelihood of mortality in some diseases, such as sepsis and pneumonia [20,21,22]. The most significant factor in this study is the use of the qSOFA score, which has not previously been applied to patients with upper GI bleeding. In terms of the size of areas under the ROC curve, the qSOFA score ranked first, with a cut-off value of 0.5. Concerning the need for intensive care, qSOFA was more useful than the GBS (0.856), the RS (0.828), and PRES (0.796), with a calcu- lated AUC of 0.921. This difference may be a result of the hemodynamic stability of the patients. Rapid hemodynamic evaluation is crucial in upper GI bleeding–in Massive bleeding, where approximately 30% of the intravascular volume is lost, the patient is considered to be in hypo- volemic shock, and this is a poor prognosis criterion [23]. The parame- ters used in the staging of hypovolemic shock include heart rate, blood pressure, respiratory rate, and change in consciousness [24]. Perhaps it is thanks to the similarities between these parameters that the qSOFA score was the best at predicting the need for intensive care in patients with upper GI bleeding. This study obtained similar results to previous studies in terms of demographic data, application complaints, and other factors. The male-female ratio was found to be 1.7:1, and the av- erage age was 63.1 +- 17.2. The most common types of bleeding were melena (72%) and hematemesis (15.2%). It was also notable that inci- dences of hematochezia were at 8.3%. Prominent chronic diseases in- cluded hypertension (43.9%) and diabetes mellitus (25%), and 71.2% of the patients had a history of drug use. It was remarkable that 4.5% of new generation anticoagulants were observed to be among the drugs that increase the propensity for bleeding. In previous studies, this rate was evaluated between 1 and 2% [25,26]. The prominent endoscopic findings were duodenal ulcers (30.3%) and gastric ulcers (24.2%). Endo- scopic adrenaline injection sclerotherapy was applied to 75% of the pa- tients, according to the endoscopy results. Transfusion was performed in 25% of the patients, and the number of transfusions was approxi- mately 2.16 +- 2. As a result, 11.4% of the patients were discharged, 77.3% were hospitalized, the intensive care admission rate was 6.8%, and the mortality rate was calculated at 4.5%. Re-bleeding was seen in 6.8% of the patients, and the average hospital stay was 5.5 days.

This data is similar to many previous studies conducted, covering a considerable number of patients, and this similarity only serves to make these results more valuable [27,28]. In this study, the qSOFA score ranked second after the GBS (0.901), with an AUC of 0.866 for mortality. The RS, which was shown to be superior in determining the

Fig. 2. Mortality.

need for transfusion in previous studies, was found to be unhelpful in this study. Concerning the need for transfusion, the GBS (AUC 0.752) was higher than that of PRES (0.650), and the qSOFA score (0.481) and the RS (0.40) were both below the threshold. In this study, the best predictor of re-bleeding was the RS, with an AUC of 0.981; PRES

was the second best predictor, followed by the qSOFA score, and the cut-off value was 7.5. In terms of in-hospital adverse events (mortality, intensive care need, recurrent bleeding, and the need for transfusion), the GBS (AUC 0.822) ranked higher than the RS (0.720), PRES (0.741), and the qSOFA score (0.597).

Fig. 3. Intensive care.

Fig. 4. The need of

1. Conclusion

Age, blood pressure, heart rate, change in consciousness, respira- tory rate, and the presence of chronic diseases provide invaluable in- formation for the prognosis of patients with upper GIS bleeding. All scoring systems were effective in predicting mortality, the need for intensive care, and recurrent bleeding. The GBS was most effective

at predicting mortality and the need for transfusion, whereas the qSOFA score proved useful in predicting the need for intensive care and the RS in predicting the risk of re-bleeding. A simpler, more ef- ficient, and more easily calculated score–the qSOFA score–can be used to evaluate the severity of patients with upper GIS bleeding by the emergency services. The innovation that distinguishes this study from other existing and previous studies, is that it shows

Fig. 5. Re-bleeding.

Fig. 6. Adverse effects.

that the qSOFA score can be used effectively to predict the need for intensive care in upper GI bleeding.

1. Limitation

One of the most important limitations of our study is that patient data were obtained retrospectively. Another important limitation is that data belonging to only one center are included in the study. The variability of parameters among the scores limits one-on-one compari- sons of scoring systems.

CRediT authorship contribution statement

Bahadir Taslidere: Conceptualization, Methodology, Software, Writing – review & editing. Ertan Sonmez: Writing – original draft. Ayse Busra Ozcan: Data curation. Liljana Mehmetaj: Visualization, In- vestigation. Elmas Biberci Keskin: Validation. Bedia Gulen: Supervision.

Declarations of Funding and Competing Interest

There are no conflicts of interest. The authors received no financial support for the research

Acknowledgements

Thanks to Fatih Usen for his support to the statistics part of the article.

References

1. Dicu D, Pop F, Ionescu D, et al. Comparison of Risk scoring systems in predicting clin- ical outcome at upper gastrointestinal bleeding patients in an emergency unit. Am J Emerg Med. 2013;31:94-9.
2. Laursen SB, Oakland K, Laine L, et al. ABC score: a new risk score that accurately pre- dicts mortality in acute upper and lower gastrointestinal bleeding: an international multicentre study. Gut. 2020. https://doi.org/10.1136/gutjnl-2019-320002 Jul 28; [e-pub].
3. Abougergi MS, Travis AC, Saltzman JR. The in-hospital mortality rate for upper GI hemorrhage has decreased over 2 decades in the United States: a nationwide anal- ysis. Gastrointest Endosc. 2015;81(4):882-8.
4. Moledina SM, Komba E. risk factors for mortality among patients admitted with upper gastrointestinal bleeding at a tertiary hospital: a prospective cohort study. BMC Gastroenterol. 2017;17:165.
5. Kim BSM, Li BT, Engel A, et al. Diagnosis of gastrointestinal bleeding: a practical guide for clinicians. World J Gastrointest Pathophysiol. 2014;5:467-78.
6. Kim MS, Choi J, Shin WC. AIMS65 scoring system is comparable to Glasgow- Blatchford score or Rockall score for prediction of clinical outcomes for non- variceal upper gastrointestinal bleeding. BMC Gastroenterol. 2019;19:136.
7. Budimir I, Gradiser M, Nikolic M, et al. Glasgow Blatchford, pre-endoscopic Rockall and AIMS65 scores show no difference in predicting rebleeding rate and mortality in variceal bleeding. Scand J Gastroenterol. 2016;51:1375-9.
8. Custovic N, Husic-Selimovic A, Srsen N, Prohic D. Comparison of Glasgow-Blatchford score and Rockall score in patients with upper gastrointestinal bleeding. Mediev Archaeol. 2020;74(4):270-4.
9. Umemura Y, Ogura H, Gando S, et al. Prognostic accuracy of quick SOFA is different according to the severity of illness in infectious patients. J Infect Chemother. 2019;25 (12):943-9.
10. Forrest JA, Finlayson ND, Shearman DJ. Endoscopy in gastrointestinal bleeding. Lan- cet. 1974;2:394-7.
11. Monteiro S, Goncalves TC, Magalhaes J, et al. Upper gastrointestinal bleeding risk scores: who, when and why? World J Gastrointest Pathophysiol. 2016;7:86-96.
12. Simon TG, Travis AC, Saltzman JR. Initial assessment and resuscitation in nonvariceal upper gastrointestinal bleeding. Gastrointest Endosc Clin N Am. 2015;25(3):429-42.
13. Budimir I, Stojsavljevic S, Barsic N, Biscanin A, Mirosevic G, Bohnec S, et al. Scoring systems for Peptic ulcer bleeding: which one to use? World J Gastroenterol. 2017; 23(41):7450-8.
14. Afessa B. Triage of patients with Acute gastrointestinal bleeding for intensive care unit admission based on risk factors for poor outcome. J Clin Gastroenterol. 2000; 30:281-5.
15. Thongbai T, Thanapirom K, Ridtitid W, et al. Factors predicting mortality of elderly patients with Acute upper gastrointestinal bleeding. Asian Biomed. 2016;10:115-22.
16. Patel V, Nicastro J. Upper gastrointestinal bleeding. Clin Colon Rectal Surg. 2020;33: 42-4.
17. Lieberman S, Valentin R, Lara L, Krempley B. Can the Glasgow Blatchford bleeding score be used as a criteria for admission to the ICU in cases of an acute gastrointes- tinal bleed? Chest. 2016;150(4):287A.
18. Blatchford O, Murray WR, Blatchford M. A risk score to predict need for treatment for upper-gastrointestinal haemorrhage. Lancet. 2000;356:1318-21.
19. Koch C, Edinger F, Fischer T, Brenck F, Hecker A, Katzer C. Comparison of qSOFA score, SOFA score, and SIRS criteria for the prediction of infection and mortality among surgical intermediate and intensive care patients. World J Emerg Surg. 2020;15(1):63 25.
20. Seymour CW, Liu VX, Iwashyna TJ, et al. Assessment of clinical criteria for sepsis: for the third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016;315(8):762-74 23.
21. Goulden R, Hoyle MC, Monis J, et al. qSOFA, SIRS and NEWS for predicting inhospital mortality and ICU admission in emergency admissions treated as Sepsis. Emerg Med

J. 2018;35(6):345-9.

1. Tokioka F, Okamoto H, Yamazaki A, Itou A, Ishida T. The prognostic performance of qSOFA for community-acquired pneumonia. J Intensive Care. 2018;6:46.
2. Cannon JW. Hemorrhagic shock. N Engl J Med. 2018;378:370-9.
3. Standl T, Annecke T, Cascorbi I, et al. The nomenclature, definition and distinction of types of shock. Dtsch Arztebl Int. 2018;115(45):757-68.
4. Park HW, Jeon SW. Clinical outcomes of patients with non-ulcer and non-variceal upper gastrointestinal bleeding: a prospective multicenter study of risk prediction using a scoring system. Dig Dis Sci. 2018;63:3253-61.
5. Gu ZC, Wei AH, Zhang C, et al. Risk of major gastrointestinal bleeding with new vs conventional oral anticoagulants: a systematic review and meta-analysis. Clin Gastroenterol Hepatol. 2020;18:792-9.
6. Stanley AJ, Laine L, Dalton HR, et al. Comparison of risk scoring systems for patients presenting with upper gastrointestinal bleeding: international multicentre prospec- tive study. BMJ. 2017;356:6432.
7. Shafaghi A, Gharibpoor F, Mahdipour Z, Samadani AA. Comparison of three risk scores to predict outcomes in upper gastrointestinal bleeding; modifying Glasgow-Blatchford with albumin. Rom J Intern Med. 2019;57(4):322-33.