Article Text
PDF
Abstract
Objective Dieulafoy’s lesions (DLs) are a rare but potentially life-threatening source of gastrointestinal (GI) haemorrhage. They are responsible for roughly 1%–6.5% of all cases of acute non-variceal GI bleeding.
Here, we describe retrospectively the clinical and endoscopic features, review the short-term and long-term outcomes of endoscopic management of bleeding DLs and we identify rate and risk factors, of recurrence and mortality in our endoscopic unit.
Design Data were collected from patients presenting with GI haemorrhagic secondary to DLs between January 2018 and August 2023. Patients’ medical records as well as endoscopic databases were retrospectively reviewed. Demographic data, risk factors, bleeding site, outcomes of endoscopy techniques, recurrence and mortality rate were taken into account.
Results Among 1170 cases of GI bleeding, we identified only seven cases involving DLs. Median age was 74 years, with a male-to-female ratio of 2.5. 75% of patients had significant comorbidities, mainly cardiovascular diseases. Only anticoagulant and antiplatelet agents were significantly associated with DLs. All patients were presented with GI bleeding as their initial symptom. The initial endoscopy led to a diagnosis in 85% of the cases. Initial haemostasis was obtained in all patients treated endoscopically. Nevertheless, the study revealed early recurrence in two out of three patients treated solely with epinephrine injection or argon plasma coagulation. In contrast, one of three patients who received combined therapy, experienced late recurrence (average follow-up of 1 year). Pathological diagnosis was necessary in one case. One patient (14%) died of haemorrhagic shock. Average length of hospital stay was 3 days.
Conclusion Although rare, DLs may be responsible for active, recurrent and unexplained GI bleeding. Thanks to the emergence of endoscopic therapies, the recurrence rate has decreased and the prognosis has highly improved. Therefore, the endoscopic approach remains the first choice to manage bleeding DLs.
- GASTROINTESTINAL BLEEDING
- ENDOSCOPIC PROCEDURES
- GASTROINTESTINAL HAEMORRHAGE
- ENDOSCOPY
Data availability statement
No data are available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Dieulafoy’s lesion (DL) is responsible for active, recurrent and unexplained non-variceal gastrointestinal bleeding.
DL is classically located in the stomach, particularly along the lesser curvature.
The gold standard treatment in endoscopic therapy with high haemostasis rate.
Selective arterial embolisation through angiography or wedge resection as the preferred surgical procedure are two management options for recurrent and uncontrolled DLs.
WHAT THIS STUDY ADDS
The majority of cases of bleeding DL are diagnosed by endoscopic techniques and, in some cases, endoscopy is non-diagnostic and no bleeding site could be identified.
Prognosis mainly depends on patient comorbidities.
Anticoagulants and antiplatelet agents emerge as significant risk factors in the occurrence of DL bleeding.
Endoscopic banding ligation or haemoclipping are the most effective and successful option of treatment of DL and epinephrine injection is a suboptimal treatment.
Combined therapy gave better results than monotherapy in terms of permanent haemostasis. However, re-bleeding can also be shown with combined therapy.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
When oesophagogastroduodenal endoscopy and colonoscopy are unremarkable in suspected haemorrhagic DL, gastroenterologists should proceed with alternative techniques (single or double balloon enteroscopies for distal small bowel/CT angiography or surgery) to identify the bleeding site.
Mechanical therapy and combined therapy should not rule out the risk of re-bleeding.
Recurrence risk factors should be sought if bleeding DL was suspected. But, their absence does not rule out the diagnosis of re-bleeding.
Introduction
Dieulafoy’s lesion (DL) can be defined as a rupture of an abnormally large submucosal artery protruding into the mucosa through a small defect in the gastrointestinal (GI) wall. It could be responsible for 2% of acute GI haemorrhages and 5.8% of non-variceal upper GI bleeding but also rare lower GI bleeding.1 2 The stomach, along the lesser curvature, is the most common site of the majority of DLs.3 4 Although rare, DLs could be an important cause of life-threatening GI bleeding especially in the elderly. In addition, the mortality rate has declined from 80% to 9% with the advent of endoscopic approaches.
During this study, we retrospectively outlined the epidemiological, clinical and endoscopic features of bleeding DL. The aim objective was to assess our institutional experience in the management of DLs and to juxtapose the resulting outcomes with existing data in the medical literature.
Design
This section outlines the experimental design, the materials used and the specific methodology adopted in conducting this study. This study was carried out in the Hepato-Gastroenterology and Digestive Endoscopy Department, University Hospital Center, spanning from January 2018 to August 2023.
Endoscopic data were retrospectively reviewed from emergent oesophagogastroduodenal endoscopy (OGD) and colonoscopy procedures conducted on 1170 patients presenting with GI bleeding. Furthermore, data were collected from two distinct digestive endoscopy units, namely the emergency unit and the interventional endoscopy unit. Procedures were carried out either by certified endoscopists or by emerging endoscopists under the guidance of an experienced team.
The seven patients who were diagnosed with DLs received intravenous proton pump inhibitors before endoscopic treatment. Thirteen patients presenting GI bleeding of unknown cause were excluded from this study.
All endoscopic examinations were performed according to international guidelines using two types of standard video endoscopes: PENTAX MEDCIAL 27″ Radiance, SC-WU27-G1522, 90X0660 and FUJIFILM 26″ Radiance, G2 HB, SC-WU26-A1511, 90XO623.
The endoscopic criteria employed for diagnosing DLs included the following: active arterial oozing, a bleeding point lacking visible underlying ulceration, micropulsatile streaming through a minute mucosal defect and visualisation of an adherent clot with a point of attachment to a tiny mucosal defect or the surrounding normal mucosa.
Various endoscopic haemostasis techniques were available; however, the choice of treatment approach was left to the discretion of the endoscopist and depended on the availability of equipment. The therapeutic methods used consisted of epinephrine injection (1:10 000 dilution) into and around the DL, argon plasma coagulation (APC) and haemoclipping.
To assess the success of therapeutic endoscopy, particular attention was given to the absence of bleeding from the DLs until the endoscope was withdrawn. Furthermore, follow-up endoscopy was conducted within a 24-hour to 48-hour timeframe following the initial procedure to confirm healing at the previous bleeding site.
Re-bleeding was defined as the presence of haemorrhagic shock accompanied by signs of ongoing blood loss following the initial transfusion and stabilisation. In cases where early recurrence (within 72 hours) was suspected, patients underwent endoscopic relook and retreatment. Surgical intervention was considered only when bleeding persisted after the second endoscopy or when the bleeding source could not be identified. To assess the possibility of late recurrence bleeding, long-term follow-up (up to 1 year) was conducted for all patients who had undergone endoscopic treatment.
Various factors were taken into consideration, including demographic data (age, gender, bleeding site and comorbidities), risk factors for GI bleeding such as medication use (non-steroidal anti-inflammatory drugs (NSAIDs), anticoagulants and antiplatelet agents), mode of presentation, laboratory results, blood transfusion requirements before and after endoscopic management, endoscopic techniques and outcomes, as well as re-bleeding and mortality rates.
Descriptive statistics were used to summarise patients’ data using Windows-based Excel software, V.2013. Categorical and continuous data were reported as percentages and medians (with IQR).
Results
Among the 1020 OGD and 150 colonoscopies performed for acute GI bleeding in our endoscopy department between January 2018 and August 2023, DLs were identified as the aetiological factor in approximately 0.6% of patients presenting with acute GI bleeding.
In the study group, seven patients were diagnosed with DLs. Among these cases, five were male (71%), while two were female (28%). The median age of the patients was 74 years, with the youngest being 50 years old and the oldest 90. All of the cases presented with significant cardiovascular risk factors. Notable comorbidities included high blood pressure in two patients, heart disease in four patients and type II diabetes in three patients. Furthermore, a combination of comorbidities was observed in two patients.
Among the seven patients, only three were not taking any medications at the time of presentation. Two patients were on direct oral anticoagulants, one due to valvular heart disease and the other for severe heart failure. A single patient was taking acenocoumarol for arrhythmia-related heart disease, and an antiplatelet agent was prescribed for the patient with aortic valvular stenosis.
Five patients presented with isolated melena, one patient experienced both haematemesis and melena, and one patient had haematemesis along with haemodynamic instability necessitating a blood transfusion. No cases of rectal bleeding were reported.
The mean haemoglobin level at presentation was 74 g/L. Blood transfusion requirements varied from 2 to 5 units of packed red blood cells (mean of 2.8 units per patient). All patients exhibited preserved biological haemostasis parameters.
Demographic and clinical data of these series are summarised in online supplemental table 1.
Supplemental material
DLs were identified in the stomach in two patients, specifically within the fundus. In two other cases, these lesions were located in the second duodenal portion. Additionally, one case was found to be situated in the bulb.
Throughout the study period, there was only one reported case of colonic DL, which was located in the right colic angle. Another case of ileal DL was also reported.
Six cases of DLs were diagnosed only by standard video endoscopes. The mean of 2.5 endoscopy sessions (range 1–4) was required for diagnosing and management of the DLs.
Different endoscopic haemostasis techniques were used. Initial haemostasis was achieved for six patients.
Epinephrine injection (1:10 000 dilution) was performed in only one patient (16%), allowing initial haemostasis. The maximal injection volume was 10 mL. However, the patient had re-bleeding (within 24 hours). Permanent haemostasis was achieved by using both APC and haemoclipping during subsequent endoscopy.
In our analysis, only two patients (33% of the cases) were managed solely with APC, without subsequent administration of epinephrine.
In one of these cases, early re-bleeding occurred, necessitating a second-look endoscopy to regain haemorrhage control. Unfortunately, the patient succumbed to haemorrhagic shock prior to endoscopic intervention.
In the remaining three patients in our study, a combination of therapeutic modalities was employed. Specifically, two cases were treated with both epinephrine injection and haemoclipping, as shown in figure 1, while the third case underwent haemoclipping followed by APC application. In these three instances, the initial interventions successfully achieved haemostasis. Nevertheless, one of these cases experienced delayed recurrent bleeding 1 year after the initial treatment.
Illustrative pictures of patient with active bleeding from fundic DL and haemostatic success after endoscopic treatment. (A) Voluminous gastric blood clot with punctiform sheet bleeding on the lesser curvature. (B) Endoscopic management with both epinephrine injection and haemoclipping.
There were no post-endoscopy complications. Table 1 summarises the earlier findings.
Endoscopic data before and after endoscopic therapies in patients with bleeding Dieulafoy’s lesions
In one case, upper endoscopy initially failed to detect any indications of haemorrhage. Subsequent colonoscopy, however, revealed the presence of blood in the ileum, although the precise origin of the bleeding remained unclear. CT angiography subsequently showed active contrast extravasation within the central section of the small bowel, devoid of any visible mass. Unfortunately, both arterial embolisation and enteroscopy were unfeasible in the emergency setting. Concurrently, the patient’s haemodynamic condition deteriorated rapidly leading to an emergent exploratory laparotomy which successfully localised the site of bleeding, a segmental resection of the small bowel was performed. Subsequent pathological analysis confirmed the presence of a DL in the proximal ileum (figures 2 and 3).
Surgical specimen showing a blood clot in the ileum.
Pathology specimen showing Dieulafoy’s lesion in the ileum.
The average length of hospital stay was 3 days. During the 1-year follow-up period, a single patient exhibited delayed recurrent bleeding attributed to DL. One patient (14% of cases), was died from bleeding-related complications.
Discussion
DL was initially reported by Gallard in 1884. Another French surgeon, Georges Dieulafoy, described this lesion in 1898 as ‘Exulceratio simplex’.3 DLs account for approximately 1.5% of all cases of acute upper GI bleeding and 3.5% of cases of jejunoileal bleeding.5 In our study, 0.6% of patients presented with acute GI bleeding secondary to DLs (seven cases). Although it is rare, DL could be responsible for severe, active and recurrent haemorrhages.6
Although old publications report a predilection towards advanced age, DLs can affect individuals of all age groups, with a median age in the sixth decade. We observed a distinct male predominance, with a sex ratio of two men to one woman.3 6 Our patients fell within the same age range as reported in previous studies. However, we noted a significant male predominance, twice that which is mostly described in the literature. Moreover, the majority of affected patients are elderly with several comorbidities. In our series, as in other studies, 85% of patients were older than 65 years.
DL is anatomically defined as a vascular malformation characterised by an aberrant, tortuous, consistently large submucosal artery that causes bleeding in the absence of ulcers, aneurysms or erosions. In the literature we reviewed, a congenital anomaly has been suggested as a cause in younger individuals, whereas an acquired malformation has been proposed for elderly.6
The presence of cardiovascular comorbidities, high blood pressure, chronic renal failure and diabetes are recognised as risk factors for the development of DLs. Several studies have suggested a correlation between bleeding DLs and chronic NSAID and anticoagulant use (in more than 40% of patients). Our findings are consistent with the role of anticoagulants and antiplatelet agents as risk factors in the occurrence of bleeding DLs. However, other authors have contested this causal association. No causal link has been established between alcohol or tobacco consumption and the occurrence of DLs.6–9
DLs are primarily found in the stomach. Up to 80% of these lesions are located 6–10 cm from the gastro-oesophageal junction, along the lesser curvature of the stomach. Elsewhere, as reported in several studies, DLs can also occur in other locations, including the duodenum, the distal stomach and the oesophagus.3 4 6 8 In our series, DLs were identified in both the stomach and the duodenum at similar rates (28% each).
Barbier et al10 reported colonic DLs as an unusual cause of lower GI bleeding in 1985. Actually, this location is extremely rare, with fewer than 30 cases documented in the literature. Baxter and Aly3 reported that colonic DLs accounted for 2% of their review of 45 case reports.
Ileal DLs represent an uncommon site for GI bleeding, comprising less than 1% of all DLs cases,3 According to the literature review, approximately 10 cases of histologically proven ileal DLs have been reported.11
Recent case reports have also described DLs occurrences in the cecum, appendix, jejunum and ano-rectum.12–14 Nikolaidis et al9 described that 43% of DLs cases were located at surgical anastomosis sites.
DLs are often asymptomatic.15 The most common clinical presentation includes melena, followed by haematemesis or haematochezia. Haemoptysis or haemorrhagic shock has also been reported. The association of haematemesis and melena was found in almost 50% of patients.6 In contrast, in our case series, 71% of patients presented with isolated melena, 14% presented only with haematemesis and 14% presented with both haematemesis and melena. Due to the severity and recurrence of GI bleeding, laboratory abnormalities were frequently observed. In our findings, 71% of our patients required transfusion of packed red blood cells, with a mean of 2.8 units.
According to the literature review, initial endoscopy has enabled a diagnosis in 65%–90% of cases. In the studies by Kasapidis et al2 and Norton et al16 an average of 1.9 or 1.3 endoscopic sessions were required for diagnosing DLs. In our study, DLs were diagnosed in 85% of patients during the initial endoscopy, with an average of 2.5 endoscopic sessions. The endoscopic criteria necessary for diagnosis included active arterial oozing, the presence of a bleeding point with no visible underlying ulceration, a protruding vessel, and visualisation of an adherent clot with attachment to a normal mucosal defect or through the surrounding normal mucosa.6 8 In the current study, 28% of patients had an adherent clot, and approximately 71% of patients presented with active bleeding, either as spurting or oozing.
Indeed, in 33% of cases, establishing the diagnosis required more than one endoscopy.8 In our study, approximately 20% of patients needed more than one endoscopy to confirm the diagnosis. Our findings were consistent with the literature.
Some unusual locations of bleeding DLs were challenging to diagnose. For instance, ileal DLs may be overlooked and inaccessible to conventional endoscopy. High volume bleeding may also be a factor that limits the sensitivity of endoscopy. Recently, single or double balloon enteroscopies identified DLs of the distal small bowel in 3.5% of patients.5 Moreover, CT angiography may be useful to determine the site of bleeding. It can be shown as contrast extravasation from an eroded tortuous artery,8 17 as in our case of ileal DL.
Pathological diagnosis, while may not be a routine practice, can become indispensable in diagnosis of DLs, especially in certain rare and challenging anatomical sites. Actually in our ileal case of DL, the diagnosis was exclusively established through histological examination.
Surgical wedge resection has been the standard treatment of DLs. Through emergent endoscopy, this option gradually replaced surgery and became first-line treatment for DLs.18 However, no consensus is recommended for endoscopic management.
Actually, the majority of data collected on treatment approaches consist of small case series or limited retrospective reports. Three main endoscopic haemostatic techniques can be used3 17: thermal (heat probe coagulation or APC, regional injection (epinephrine or norepinephrine) and sclerotherapy, mechanical therapy (endoscopic band ligation (EBL) and haemoclipping).
Although the results of endoscopic haemostatic therapies were variable, the rate of permanent haemostasis was high, reaching 90%. In a retrospective study, APC was found to be very effective for management of bleeding DLs.19 In our study, a re-bleeding was observed in one patient treated with APC, while permanent haemostasis was achieved in the other. Mechanical haemostasis provided by EBL or haemoclipping is the most effective option in the treatment of DLs with a high success rates (95%) and low risk of recurrence <10%.20 21 Due to the risk of re-bleeding, epinephrine injection is considered as a suboptimal treatment and it is not recommended to be used alone.8 These results are consistent with data achieved by this study. The use of combined therapy gave better results than monotherapy in terms of permanent haemostasis and low rate of re-bleeding.8 22 By contrast, in this study, one patient treated by combined therapy including epinephrine and haemoclipping had a recurrent bleeding during the follow-up period.
Comparative results of endoscopic therapies from retrospective studies are summarised in table 2.
Summary of findings from the literature on endoscopic therapy of patients with Dieulafoy’s lesions
Based on the current literature, the rate of re-bleeding is 9%–40% and increases with endoscopic monotherapy.8 The overall risk of recurrence in the short-term (<72 hours) is around 10%.23 In instances of recurrent and uncontrolled bleeding, two management options have been documented for DLs6 8: selective arterial embolisation through angiography or wedge resection as the preferred surgical procedure.
Thanks to emergent endoscopy, the mortality of DLs has significantly decreased from 80% to 8.6%. In the current study, one patient died from haemorrhagic shock (14%).
This study is subject to several inherent limitations. First, the relatively rare incidence of DLs has resulted in a limited sample size, which may compromise the ability to establish statistical significance for the findings. Furthermore, being a retrospective investigation, there is a potential for inaccuracies in the recorded medical histories and risk factors.
Notwithstanding these constraints, the study reveals a robust association between cardiovascular diseases and episodes of bleeding attributed to DLs. Additionally, the use of anticoagulants and antiplatelet agents emerges as significant risk factors in the development of DLs.
Conclusion
While uncommon, DL is unequivocally acknowledged as a potentially life-threatening source of acute GI haemorrhage, particularly among the elderly. Timely diagnosis and prompt endoscopic management, especially in elderly individuals with multiple risk factors, can be invaluable and lifesaving.
Haemoclipping and combined therapeutic modalities are emerging as the most efficacious approaches, demonstrating superior haemostasis rates when compared with injection therapy. Nevertheless, to establish the optimal endoscopic treatment for DLs, further comprehensive studies will be imperative in the future.
Data availability statement
No data are available.
Ethics statements
Patient consent for publication
Ethics approval
This retrospective analysis was conducted according to the ethical principles of the Declaration of Helsinki and local legislation, informed consent was obtained from the seven participants for this retrospective study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
X @N/A
Contributors BA: manuscript writing and analysis. GK: manuscript correction. AZ: verification of the overall reproducibility of the results. HK: evolution of the general aims and objectives of research. OE: methodology development. IZ: data visualisation and presentation. BA and IZ confirm and accept full responsibility for the overall content of the work as guarantors.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.