a b s t r a c t

Objective: The assessment of the severity of upper gastrointestinal hemorrhage in emergency department (ED) patients is difficult to assess with commonly available diagnostic tools. Small studies have shown that video cap- sule endoscopy (VCE) is a promising risk-stratification method and may be better than current clinical decision rules such as the Rockall Score and the Glasgow Blatchford score. This review aims to assess the accuracy of VCE to detect active upper gastrointestinal hemorrhage compared to a reference standard.

Methods: The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodology was used to perform a review of studies that have measured the diagnostic accuracy of VCE. Studies were included if they measured ED use of VCE for upper GI hemorrhage as compared to a reference standard of an esophagogastroduodenoscopy . A meta-analysis was performed on select patients using a fixed effects and random-effects model to determine the primary outcome of Diagnostic test accuracy.

Results: 40 studies were screened for eligibility and five studies representing 193 patients met the inclusion and exclusion criteria. All patients received both a VCE and an EGD. The sensitivity and specificity of VCE were 0.724 and 0.748, respectively. The diagnostic odds ratio was 6.29 (95% CI: 3.23-12.25) and the summary receiver op- erating characteristic curve was 0.782.

Conclusions: VCE demonstrated high accuracy for detecting upper GI hemorrhage in this meta-analysis of existing studies. In light of the potential advantages of VCE in the ED, further research is warranted to further establish its role.

(C) 2020

Introduction

Background

? This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

* Corresponding author at: The Department of Emergency Medicine, 2120 L Street NW, Suite 450, Washington, DC 20037, USA.

E-mail addresses: [email protected] (N. Shah), [email protected] (C. Chen), [email protected] (N. Montano), [email protected] (N.T. Gentile), [email protected] (A.T. Limkakeng), [email protected] (A.B. Kumar), [email protected] (Y. Ma), [email protected] (A.C. Meltzer).

Gastrointestinal (GI) hemorrhage is a common presentation in US emergency departments (EDs), accounting for about 243,675 visits in 2014 [1]. For upper GI hemorrhage, the standard of care is inpa- tient hospital admission and an esophagogastroduodenoscopy (EGD). More recently, the use of Video capsule endoscopy (VCE) has been studied as a method of risk stratification in the ED with the goal of expediting diagnosis, decreasing hospital admissions, and improving efficiency [2].

https://doi.org/10.1016/j.ajem.2020.03.008 0735-6757/(C) 2020

Importance

ED risk stratification of upper GI hemorrhage is typically per- formed by clinical decision rules such as the Glasgow-Blatchford score (GBS) or nasogastric lavage. Both of these approaches have sig- nificant limitations. The GBS only identifies a small number of pa- tients who can safely be discharged while nasogastric lavage is uncomfortable to patients, may have complications, and has a low sensitivity and specificity for active bleeding [3,4]. Consequently, the vast majority of patients need hospital admission and inpatient EGD due to the risk of re-bleeding and mortality in the presence of an active bleed [5]. Once admitted, the average length of stay for upper GI hemorrhage is 2.9 days and the average hospital charge is

$37,232 [1]. VCE is a Novel method of risk stratification that has the potential to provide real-time imaging of the upper GI tract, reduce hospital admissions and improve the accuracy of ED-based risk strat- ification. Models that have compared VCE to GBS have demonstrated that VCE may be cost-effective [2].

Goals of investigation

The purpose of this review is to synthesize the results of several small studies that examined the accuracy of VCE compared to a gold standard of EGD to detect upper GI bleeds in the ED. Our goal was to provide Diagnostic test characteristics including area under the curve (AUC) and diagnostic odds ratio (DOR) for VCE to gauge the potential of VCE as an ED risk stratification tool and to inform fu- ture studies.

Methods

Study design and setting

This systematic review was conducted following the PRISMA-DTA Checklist [6] to evaluate the diagnostic accuracy of VCE in detecting upper GI bleed in the ED. The pre-specified search protocol was developed with a research team, including a medical research librarian. A protocol for this review is registered with PROSPERO (ID: CRD42018104118). In order to be included, studies must have recorded the results of the index test (VCE) and the reference standard (EGD) for each participant. Only studies conducted in the ED with patients pre- senting specifically for upper GI bleed were included. Appendix 1 in- cludes the complete search strategy with the inclusion and exclusion criteria.

Search and study selection

A systematic literature review using PubMed, Scopus and Cochrane CENTRAL was conducted using the following keywords: video capsule endoscopy, capsule endoscopy, PillCam, wireless capsule, emergency department, emergency medicine, emergency triage, emergency room, upper gastrointestinal hemorrhage and Upper gastrointestinal bleeding. No language or publication date restrictions were imposed. methodological quality“>Two reviewers independently conducted the searches using the pre- determined search strategy. Only full-text articles were considered; active studies, abstracts, and unpublished data were excluded. In Phase 1, titles and abstracts were assessed for eligibility by two re- viewers independently. Studies meeting the eligibility criteria were then reviewed in full-text by both reviewers independently in Phase 2. Full-text of the articles was screened, and studies were se- lected based on the Inclusion/exclusion criteria. Any disagreements were resolved by consensus.

Data extraction

Data were extracted by two reviewers and any disagreements were resolved by consensus. For each study, the following variables were ex- tracted: type of study, patient characteristics (such as age and gender), type of PillCam, sample size, presenting signs, average blood pressure, and average time to EGD (Table 1). For the Gralnek study, data was ex- tracted from Table 3 by comparing the diagnosis made by the VCE and EGD modality [7]. If a bleeding lesion was detected, it was considered a positive result; whereas a normal or non-bleeding lesion was consid- ered a negative result.

Quality assessment for bias and applicability

The quality of the studies in the review was assessed by one re- viewer using the revised Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) [8]. A study was regarded as low risk-of-bias, if all the answers within the domain were recorded as yes; whereas a study was regarded as high risk-of-bias, if at least one answer in the do- main was no. Unclear risk was assigned to studies if any of the answers in the domain were unclear.

Statistical analysis and quantitative synthesis

The pooled sensitivity and specificity, diagnostic odds ratio and the Area Under the Receiver Operating Characteristic Curve with corresponding 95% confidence intervals (CI) were calculated based on the hierarchical summary receiver operating characteristic (HSROC) model [9]. The main results were summarized and pre- sented by a summary receiver operating characteristic (SROC) curve. Heterogeneity due to threshold was assessed through the spearman correlation coefficient between the logarithms of sensitiv- ity and (1-specificity); while heterogeneity caused by non-threshold was evaluated by the Cochran Q statistic and quantified by the con- sistency index I², where a higher value of I² indicates a higher level of heterogeneity. We used Deek’s effective sample size funnel plot to test for potential publication bias, where a p-value of b0.10 is sug- gestive of significant publication bias [10]. To identify possible sources of heterogeneity, we performed meta-regression by adding covariates to a bivariate random effects model [11]. A p-value of b0.05 was considered as statistically significant. The Midas and Metandi module for Stata 15.0 (Stata Corporation, College Station, TX, USA), and the Madas package in R (version 3.5.2) were used to conduct meta-analysis and produce all graphical displays.

Results

Literature search

The literature search yielded a total of 40 articles, of which 10 were excluded as duplicates. Title and abstract review were conducted on the remaining 30 articles. Upon full text review, 9 studies were excluded for the following reasons: study did not occur in the ED (n = 1), study was not specific to upper GI bleeding (n = 3), study did not include an accu- racy comparison (n = 1), editorial (n = 1), review article (n = 1), and same patient population (n = 2). Fig. 1 depicts the complete flowchart for the literature search.

Methodological quality

The four domains of the QUADAS-2 tool were used to evaluate the studies in the review (Fig. 2) [8]. The risk-of-bias criterion for flow and timing was considered moderate due to variable time from VCE until gold standard of EGD. In the patient selection criterion, risk-of-bias was moderate because patients were enrolled on a con- venience basis rather than consecutively or randomly. In addition,

Table 1

Study characteristics.

Author, year	Study design	Average age	Sex	Average BP	Presenting signs	Type of PillCam	Sample size (VCE & EGD)	Average time to EGD (h)
Gralnek, 2013	Prospective cohort study	58.3 +- 19	M: 30; F:16	137/72	Melena alone (n = 32) Hematemesis alone (n = 10)	Modified PillCam ESO 2	46	12-24
Meltzer, 2013	Prospective	52	M: 9; F: 10	Median SBP: 120	Melena and hematemesis (n = 7) Melena (n = 11)	PillCam ESO 2	19	12-24
Chandran,	Cohort Study Diagnostic,	Median age:	M/F —	Median SBP: 124	Coffee ground emesis (n = 12) Both CG and melena (n = 2) Hematemesis (n = 22)	PillCam ESO	83	12-24
2013	non-randomized, single-blind study	65	3.6/1		Melena (n = 42) Hematemesis and melena (n = 19)
Rubin, 2011 Sung, 2016	Prospective randomized controlled trial Prospective	61.2 +- 24.4 55.2 +- 18.7	M: 8; F: 4 M: 21;	MAP: 78.47 Systolic BP: 121.75 SBP: 132.4	Hematemesis (n = 12) Melena (n = 17) Presyncope/syncope (n = 19) “Coffee ground” vomiting	PillCam ESO 2 PillCam ESO 2	11 34	b12 N24
	randomized controlled trial		F:13		(n = 15) Melena (n = 51) Both CG and melena (n = 2)

there was unclear risk of bias for reference standard due the fact that EGD interpretations might not have been strictly blinded from VCE interpretation. In terms of applicability concerns, the risk-of-bias cri- terion for all domains for all five studies was low risk. Overall, the quality of the studies included in the review was considered to be fair.

Diagnostic accuracy

A total of 5 studies reporting diagnostic accuracy measures matched the inclusion criteria and thus were included in the meta-analysis. The paired forest plot presented in Fig. 3 showed that the sensitivity ranged

from 0.62-1.00 and the specificity ranged from 0.64-1.00. The pooled sensitivity of 5 studies was 0.78 (95% CI = 0.58-0.90), and the pooled

specificity was 0.78 (95% CI = 0.68-0.86). The results in Fig. 4 and Fig. 5 revealed that the overall DOR of VCE was 12.62 (95% CI = 3.90-40.85) and the AUC was 0.82 (0.68-0.90), indicating a moderately high level of overall accuracy.

The Spearman correlation coefficient of sensitivity and false positive rate (? = 0.1, p = 0.95 N 0.05) and the visual inspection of SROC curve suggested that the threshold effect did not contribute to the observed heterogeneity. Meanwhile, significant heterogeneity between studies was observed in the overall sensitivity (Q = 8,45, I² = 52.65), specificity (Q = 4.62, I² = 13.33), and DOR (Q = 20.59, I² = 80.57). Therefore, we

Fig. 1. PRISMA flow diagram.

Fig. 2. Risk of bias and applicability concerns overview using QUADAS-2.

concluded that only non-threshold effect accounted for the detected heterogeneity.

We subsequently conducted meta-regression to explore potential heterogeneity using covariates such as mean age, proportion of males, and patient sample size. The results of meta-regression presented in Table 2 revealed that among these covariates, sample size significantly accounted for the heterogeneity of sensitivity (p = 0.037). Proportion of males (p = 0.62) and mean age (p = 0.084) appeared to be possible

contributors to the between study heterogeneity, although not significant.

Deek’s funnel plot and asymmetry test displayed in Fig. 6 illustrated

that there was direct evidence for publication bias (p = 0.052). The shape of the funnel plot suggested that smaller studies having higher in- verse square root of effective sample size tended to have higher diag- nostic odds ratio than larger studies. We performed a sensitivity analysis by excluding one study at each time, and recalculated DOR

Fig. 3. Paired forest plots of sensitivity and specificity.

Fig. 4. The forest plot of diagnostic odds ratio.

and AUC. The DOR ranged from 8.13 to 19.85, and the AUC ranged from

0.75 to 0.88 (Table 3). When a study by Rubin et al. [12] was excluded from meta-analysis, DOR decreased from 12.62 to 8.13 and I² decreased from 80.57% to 51.14%. However, the overall accuracy of VCE did not seem to vary dramatically with the exclusion and remained at a good level. Therefore, the results suggested that our analysis was stable, and indicated that VCE could offer a satisfactory degree of accuracy for the diagnosis of upper GI hemorrhage.

Time to EGD

Time to EGD after VCE was relatively consistent across all studies. EGD was performed b12 h from VCE in one study, performed between 12 and 24 h in three studies and performed N24 h after VCE in one study [7,12-15]. Furthermore, these EGDs took place in the endoscopy suite and not in the ED itself.

Discussion

This is the first meta-analysis to assess the accuracy of VCE and pro- vide the aggregate sensitivity and specificity of VCE in detecting upper GI bleeding in the ED setting. Our analysis shows that VCE has a high di- agnostic accuracy with a diagnostic odds ratio of 12.62 and the AUC of

0.82. To put the numbers into context, nasogastric lavage, an existing risk stratification tool in the ED, demonstrates lower odds ratio of 1.44 to 4.74 [16].

If shown to be sufficiently accurate, there are many potential advan- tages of VCE over current standard of care including patient comfort, avoidance of sedation, radiation-free, operator-independent adminis- tration, and reduced cost [2]. Patients who have had both EGD and VCE are more willing to undergo VCE again versus EGD, 84.1% versus 55%, respectively [7]. If VCE can effectively increase the ED discharge rates of low-risk patients thereby reducing hospital admissions, the test becomes very cost effective [2]. Based on a prior cost analysis study, the use of capsule endoscopy in low-risk and medium-risk

Table 3

Sensitivity analysis

Study DOR 95%CI AUC Heterogeneity

				Q	I2 (%)	p
Overall	12.62	3.90-40.85	0.82	20.59	80.57	0.00
Exclude Gralnek, 2013	16.54	2.97-92.01	0.86	50.64	94.05	0.00
Exclude Meltzer, 2013	13.95	3.32-58.54	0.85	40.93	92.67	0.00
Exclude Chandran, 2013	19.85	6.71-58.77	0.88	7.07	57.55	0.07
Exclude Rubin, 2011	8.13	3.55-18.61	0.80	6.14	51.14	0.10
Exclude Sung, 2016	9.91	3.02-32.53	0.75	11.12	73.01	0.01

Fig. 5. Hierarchical summary receiver operating characteristic plot.

Table 2

The results of meta-regression

Covariate	Sensitivity				Specificity
	Coef	SE	p		Coef	SE	p
Mean age	-0.092	0.053	0.084	-0.005		0.050	0.918
Proportion of male	-4.639	2.481	0.062	-0.866		2.299	0.706
Sample size	-0.020	0.010	0.037	0.001		0.010	0.894

patients is highly cost-effective compared to admission [2]. The global fee for capsule endoscopy (CPT 91111) is roughly $1000 including inter- pretation (www.cms.org). In comparison, patients presenting with UGIB who required hospital admission stayed for an average of 4 days, accumulating to $23,549 [2].

The strength of this review is that all included studies used a similar index test and reference standard. Four of the included studies used the second-generation PillCam ESO 2, whereas one study used the first gen- eration PillCam ESO; both possess similar technology. In addition, in all included studies, the index test was read by an expert in capsule endos- copy or a board-certified gastroenterologist. Furthermore, the index test was compared to the same reference standard, EGD, in all studies for all patients, thereby reducing risk of verification bias.

A limitation of this is the low number of studies included and the fair quality of these studies. Of the studies included, the sample size of three studies were very small (b35 patients). Subgroup analysis was not con- ducted due to a small number of studies. In addition, in general, the ref- erence standard (EGD) was not performed immediately after the index test (VCE) which may have the effect of reducing the accuracy of the VCE due to the dynamic nature of gastrointestinal lesions [17]. For the Gralnek study, data was extracted by assessing the diagnosis made by both the VCE and EGD modality but the data for sensitivity and specific- ity was not provided [7].

In addition to conducting randomized trials with larger sample sizes, future studies need to consider patient-centered outcomes to deter- mine how the use of VCE in the ED improves quality and impacts cost- effectiveness. Additionally, as the technology improves, newer features such as magnetic control will allow operators to move the capsule to a certain area of the gastrointestinal tract [18-20]. Recent literature has shown how capsules can more quickly and accurately diagnose bleed- ing lesions for inpatients with occult bleeding and how associated soft- ware has developed sophisticated means of blood detection [22].

Fig. 6. Deek’s effective sample size funnel plot for publication bias.

Finally, longer battery life capsule may be necessary to provide informa- tion regarding a potential source of bleeding in the small bowel and right colon in addition to the upper GI tract in patients with melena but no hematemesis. While capsule retention is a possibility but an un- likely side effect, finding the capsule in stool is not considered a compli- cation nor is it required for the patient. Moreover, recent advances have developed a patency capsule, in which the capsule dissolves if retained in the bowel for N48-100 h [23].

In conclusion, this systematic review and meta-analysis shows that VCE has promising diagnostic accuracy in detecting upper GI bleed for adults presenting in the ED. Given the accurate test char- acteristics, VCE has the potential to serve as a risk stratification tool to rule out UGIB. Further research on VCE should be conducted using randomizing trials with larger sample sizes to validate the modality’s diagnostic accuracy and ability to affect patient reported outcomes.

Author contribution

Dr. Meltzer serves on advisory board for AnX Robotics and has a re- search grant from Medtronic, manufacturer of the PillCam. No other au- thors have a conflict to report.

Dr. Cave has received research support from Olympus, manufacturer

of ENDOCAPSULE and Medtronic, manufacturer of the PillCam.

Dr. Limkakeng has received research support from Roche Diagnos- tics, Inc., Abbott Laboratories, Siemens Healthcare Diagnostics, Hospital Quality Foundation, Bristol Myers Squibb, Ischemia Care, LTD., GE and AstraZeneca.

ACM and NS contributed to study concept and design and acquisi- tion of the data.

ACM, YM, CC and NS contributed to analysis and interpretation of the data.

YM and CC contributed statistical expertise.

NS, CC, NM, DC, RS, YM and ACM contributed to drafting and critical revision of the manuscript.

ACM takes responsibility for the paper as a whole.

Appendix 1

Question criteria

• P (patient characteristics) — adults presenting to ER
• P (patient presentation) — upper GI bleed
• P (prior testing) — none
• I (index test) — VCE
• C (comparator tests) — EGD
• P (purpose) — triage
• T (target disorder) — upper GI bleed
• R (reference standard) — EGD

Databases to be searched

• PubMed (MESH Headings and Text)
• Scopus
• Cochrane CENTRAL

Inclusion criteria

• randomized control trials where patients are randomized into VCE or another Diagnostic modality
• Studies reporting comparative results between VCE and other diag- nostic modalities
• Systematic reviews and meta-analysis
• Case studies
• Cohort studies

Exclusion criteria

• Studies in non-humans
• pediatric studies
• Editorials
• Opinion papers
• Studies not specific to upper GI hemorrhage
• Studies not taking place in the ER
• Non-comparative observation studies
• Abstract only

Search date: December 24th, 2018 PubMed search

(((“Emergency Service, Hospital”[Mesh] OR tertiary care OR emer- gency medicine OR emergency department OR emergency triage OR emergency room)) AND (((“Upper Gastrointestinal Tract”[Mesh]) AND “Hemorrhage”[Mesh]) OR upper gastrointestinal bleeding OR upper gastrointestinal hemorrhage)) AND (“Capsule Endoscopy”[Mesh] OR Video Capsule Endoscopy OR Pillcam OR Wireless Capsule).

Scopus

TITLE-ABS-KEY (“Video Capsule Endoscopy” OR “Capsule Endos- copy” OR “Pillcam” OR “Wireless Capsule”) AND TITLE-ABS-KEY (“Emergency Department” OR “Emergency Medicine” OR “Emergency Triage” OR “Emergency Room”) AND TITLE-ABS-KEY (“Upper Gastroin- testinal hemorrhage” OR “upper gastrointestinal Bleeding”)).

Cochrane CENTRAL

Capsule Endoscopy OR Video Capsule Endoscopy OR PillCam OR Wireless Capsule in Title Abstract Keyword AND Emergency Depart- ment OR emergency medicine OR emergency Triage OR Emergency Room in Title Abstract Keyword AND Upper Gastrointestinal bleeding OR Upper gastrointestinal hemorrhage in Title Abstract Keyword (Word variations have been searched).

References

1. HCUPnet: a tool for identifying, tracking, and analyzing national hospital statistics. https://hcupnet.ahrq.gov/#query/eyJBTkFMWVNJU19UWVBFIjpbIkFUX00iXSwiT1V UQ09NRV9NRUFTVVJFUyI6WyJPTV9OVU1CRVIiLCJPTV9SQVRFIiwiT01fTG9TIiwiT0 1fSENIIl0sIllFQVJTIjpbIllSXzIwMTQiXSwiUFJJTkNJUExFX09SX0FMTCI6WyJQQV9QU

   klOQ0lQTEUiXSwiQ0FURUdPUklaQVRJT05fVFlQRSI6WyJ, Accessed date: 27 June 2018.

   Meltzer AC, Ward MJ, Gralnek IM, Pines JM. The cost-effectiveness analysis of video capsule endoscopy compared to other strategies to manage acute upper gastrointes- tinal hemorrhage in the ED. Am J Emerg Med 2014;32(8):823-32. https://doi.org/ 10.1016/j.ajem.2013.11.012.

2. Pitera A, Sarko J. Just say no: gastric aspiration and lavage rarely provide benefit. Ann Emerg Med 2010;55(4):365-6. https://doi.org/10.1016/j.annemergmed.2009.10. 016.
3. Gutkin E, Shalomov A, Hussain SA, et al. Pillcam ESO (R) is more accurate than clinical scoring systems in risk stratifying emergency room patients with acute upper gas- trointestinal bleeding. Therap Adv Gastroenterol 2013;6(3):193-8. https://doi.org/ 10.1177/1756283X13481020.
4. Chiu Philip WY, Ng EKW. Predicting poor outcome from Acute upper gastrointestinal hemorrhage. Gastroenterol Clin NA 2018;38(2):215-30. https://doi.org/10.1016/j. gtc.2009.03.009.
5. MF McInnes, Moher D, BD Thombs, TA McGrath, PM Bossuyt, PRISMA-DTA Group. Preferred reporting items for a systematic review and meta-analysis of diagnostic test accuracy studies: the PRISMA-DTA statement. JAMA 2018;319(4):388-96. https://doi.org/10.1001/jama.2017.19163.
6. Gralnek I, Ching J, Maza I, et al. Capsule endoscopy in acute upper gastrointestinal hemorrhage: a prospective cohort study. Endoscopy 2012;45(1):12-9. https://doi. org/10.1055/s-0032-1325933.
7. PF W, AS R, ME W, et al. Quadas-2: a revised tool for the quality assessment of diag- nostic accuracy studies. Ann Intern Med 2011;155(8):529-36. https://doi.org/10. 7326/0003-4819-155-8-201110180-00009.
8. Rutter CM, Gatsonis CA. A hierarchical regression approach to meta-analysis of diag- nostic test accuracy evaluations. Stat Med 2001;20(19):2865-84.
9. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58(9):882-93.
10. Reitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate analysis of sensitivity and specificity produces informative summary measures in di- agnostic reviews. J Clin Epidemiol 2005;58(10):982-90.
11. Rubin M, Hussain SA, Shalomov A, Cortes RA, Smith MS, Kim SH. Live view video capsule endoscopy enables risk stratification of patients with acute upper GI bleed- ing in the emergency room: a pilot study. Dig Dis Sci 2011;56(3):786-91. https:// doi.org/10.1007/s10620-010-1336-9.
12. Meltzer AC, Pinchbeck C, Burnett S, et al. Emergency physicians accurately interpret video capsule endoscopy findings in suspected upper gastrointestinal hemorrhage: a video survey. In: Lewis L, editor. Acad Emerg Med, 20(7). ; 2013. p. 711-5. https:// doi.org/10.1111/acem.12165.
13. Chandran S, Testro A, Urquhart P, et al. Risk stratification of upper GI bleeding with an esophageal capsule. Gastrointest Endosc 2013;77(6):891-8. https://doi.org/10. 1016/j.gie.2013.01.003.
14. Sung JJY, Tang RSY, Ching JYL, Rainer TH, Lau JYW. Use of capsule endoscopy in the emergency department as a triage of patients with GI bleeding. Gastrointest Endosc 2016;84(6):907-13. https://doi.org/10.1016/j.gie.2016.04.043.
15. Palamidessi N. nasogastric aspiration and lavage in emergency department patients with hematochezia or melena without hematemesis. Acad Emerg Med 2010;17(2): 1069-6563.
16. Meltzer AC, Ali MA, Kresiberg RB, et al. Video capsule endoscopy in the emergency department: a prospective study of acute upper gastrointestinal hemorrhage. Ann Emerg Med 2013;61(4):438-443.e1. https://doi.org/10.1016/j.annemergmed.2012. 11.008.
17. Liao Z, Hou X, Lin-Hu EQ, et al. Accuracy of magnetically controlled capsule endos- copy, compared with conventional gastroscopy, in detection of gastric diseases. Clin Gastroenterol Hepatol 2016. https://doi.org/10.1016/j.cgh.2016.05.013.
18. Liao Z, Zou W, Li Z. Clinical application of magnetically controlled capsule gastros- copy in gastric disease diagnosis: recent advances. Sci China Life Sci 2018;61(11): 1304-9.
19. Spada C, Hassan C, Costamagna G. Magnetically controlled capsule endoscopy for the evaluation of the stomach. Are we ready for this? Dig Liver Dis 2018. https://doi.org/ 10.1016/j.dld.2018.07.039.
20. Han S, Fahed J, Cave DR. Suspected blood indicator to identify active gastrointestinal bleeding: a prospective validation. Gastroenterol Res 2018. https://doi.org/10. 14740/gr949w.
21. Rondonotti E, Herrerias JM, Pennazio M, Caunedo A, Mascarenhas-Saraiva M, De Franchis R. Complications, limitations, and failures of capsule endoscopy: a review of 733 cases. Gastrointest Endosc 2005;62(5):712-6. https://doi.org/10.1016/j.gie. 2005.05.002.