Endoscopic Diagnosis and Treatment of Colonic Diverticular Bleeding

INTRODUCTION

Background

In adult patients who present to the hospital with severe hematochezia (e.g. red blood or clots per rectum), the most common colon diagnosis reported is diverticular hemorrhage (1,2). Hematochezia or lower gastrointestinal (LGI) bleeding is reported to have an annual incidence of about 20 cases/100,000 population, but with a much higher rate in the elderly (3). The majority of the patients with severe hematochezia from colon sources are older than 70 years and present to the hospital with painless bleeding, anemia with a decrease in hemoglobin from baseline, but usually without syncope or major changes in orthostatic pulse or systolic blood pressure (1,2). The most common colonic diagnoses by urgent colonoscopy among 882 patients with severe LGI bleeding reported in descending order are diverticular (TIC) hemorrhage (32.3%), ischemic colitis (11.2%), internal hemorrhoids (10.4%), post- polypectomy induced ulcer hemorrhage (9.1%), rectal ulcers (8.2%), and other forms of colitis (7.1%). Refer to table 1 for further details (4).

Table 1: Eight Most Common Colonic Sources of Severe Hematochezia (882 cases) (Expressed as percent of colonic sources).

Diverticulosis Def is definitive and Pres is presumptive. UC is ulcerative colitis. C. dif is Clostridium dificile. XRT is radiation telangiectasia. Colon CA is cancer. See reference 4

Diverticulosis (Def/Pres)	32.3%
Ischemia	11.2%
Internal hemorrhoids	10.4%
Post Polypectomy Ulcer	9.1%
Rectal ulcers	8.2%
Colitis (UC, C. dif, Crohn’s, other)	7.1%
Colon angiomas/XRT	6.8%
Colon CA or Polyps	6.0%

The current endoscopic diagnosis and treatment of documented, acute colon diverticular hemorrhage is the focus of this report. There are several reasons for this emphasis. The first is because TIC hemorrhage is the most common diagnosis in older adult patients who are hospitalized with an acute severe LGI hemorrhage. The second is that urgent colonoscopy for diagnosis and hemostasis is feasible, safe and effective when performed by experienced endoscopists in the US or elsewhere (1,2,4-6). Clinical outcomes have been improved with this management. Third, in referral hospitals which have colonoscopists with expertise in performing urgent colonoscopy and diagnosing diverticular hemorrhage, this approach has replaced emergency angiography and surgery. That is because it is more often available, diagnostic, and safer and more cost effective than non-endoscopic treatments, especially for elderly patients (5,6). Part of the reason for this change in usual treatment is that colonoscopic hemostasis itself has improved over the last 3 decades from treatment with less effective injection of epinephrine (7) to use of more effective hemostasis techniques including thermal coagulation, rubber band ligation, through-the-scope hemoclips, combination treatments, and large over-the-scope hemoclips targeted at the bleeding site (4-16).

The diagnosis of definite diverticular hemorrhage has also improved. Similar to peptic ulcer bleeding (PUB) where stigmata of recent hemorrhage (SRH) have been utilized to risk stratify for rebleeding (17), active bleeding and other non-bleeding SRH have been applied for diagnosis, risk stratification, and treatment of definitive TIC hemorrhage (2,4,5,18-21). To improve risk stratification and treatment of definitive TIC hemorrhage beyond utilizing visual SRH alone, Doppler endoscopic probe (DEP) monitoring of arterial blood flow underneath SRH has also been utilized (18,20,22). DEP has helped define the natural history of definitive diverticular hemorrhage which has led to recognition that all TIC SRH have a high risk for rebleeding. By expanding TIC SRH to include flat spots and blood flow detection by DEP, there is an overall increase in the rate of diagnosis of definitive TIC hemorrhage on urgent colonoscopy (20). DEP monitoring of arterial blood flow after colonoscopic hemostasis has also resulted in higher rates of definitive hemostasis when there is obliteration of arterial blood flow underneath TIC SRH, similar to reports for non-variceal UGI lesions (22). These technical advances and changes have contributed to an evolution in patient management, colonoscopic diagnosis, and colon hemostasis with subsequent improvements in clinical outcomes. That is when clinical results of current colonoscopic hemostasis are compared with medical management, angiography and surgery for treatment of acute diverticular hemorrhage. These results will be described and discussed in detail in this manuscript.

The specific aims of this report all relate to acute TIC hemorrhage and include: 1) to present the rationale, methods, and results for early colonoscopic diagnosis and treatment, 2) to review the short term natural history of definitive TIC hemorrhage based upon SRH and DEP monitoring, 3) to describe short and long term rebleeding rates after colonoscopic hemostasis in comparison with medical treatment, angiographic embolization, and surgery, 4) to discuss a new approach for earlier diagnosis of definitive TIC hemorrhage in patients with severe LGI bleeding, 5) to describe new treatment techniques to improve colonoscopic hemostasis so that TIC rebleeding can be reduced when compared to older colonoscopic treatments, and 6) to provide recommendations about salvage treatments for severe or recurrent TIC bleeding, including angiographic embolization, surgery, and repeat colonoscopic hemostasis by an expert endoscopist.

Definitions of Diverticular Hemorrhage

The classification of colon diverticulosis as the cause of severe LGI hemorrhage is important to put into context with the different ways “diverticular hemorrhage” is defined and reported elsewhere. Elderly patients commonly have colon diverticulosis but most do not develop diverticular hemorrhage. Prior to presenting to the hospital with an acute LGI bleed, many elderly patients have colon diverticula diagnosed by previous colonoscopy, barium enema, computerized tomography (CT), or magnetic resonance imaging (MRI). After discharge, International Classification of Diseases (ICD) coding of patients with LGI hemorrhage and diverticulosis is often not classified or reported accurately. For example, patients who stop bleeding and do not have a GI evaluation with colonoscopy or have one that is non-diagnostic are usually coded as diverticular hemorrhage for billing purposes. However, such ICD coding is neither sufficiently accurate for classification of true diverticular hemorrhage (e.g. as definitive or presumptive diverticular hemorrhage or incidental diverticulosis, as detailed below) nor for reporting these ICD based results in evidence-based research.

A more accurate classification for colon diverticulosis as the cause of an acute severe LGI hemorrhage is detailed Box 1 (2). These diagnoses are based upon whether or not there are colonic diverticula with SRH or other lesions that are found as the cause of the GI hemorrhage. This classification is highly recommended for investigators and colonoscopists who report about diagnosis, management, and clinical outcomes of patients with an acute TIC hemorrhage. If utilized, this TIC classification as it relates to an acute LGI hemorrhage will facilitate communication, interpretation, and comparisons of diagnostic findings and clinical outcomes of different subgroups of patients with colon diverticulosis and acute LGI bleeding (2).

Box 1. Classification of Diverticulosis as Cause * (or not) of Severe Hematochezia.

*Definitive diverticular bleed – stigmata of hemorrhage on a colon TIC found on urgent colonoscopy &/or surgery; or active bleeding on RBC scan or angiogram confirmed to be diverticulosis on colonoscopy.

*Presumptive diverticular bleed - diverticulosis without stigmata & no other bleeding lesions found by colonoscopy, anoscopy, enteroscopy, & capsule endoscopy.

Incidental diverticulosis - diverticulosis present but another source of bleeding is identified in the colon, anorectum, foregut , or small intestine.

Legend* These are true diverticular hemorrhage. See reference 2

METHODS

Classification of Diverticular Hemorrhage Based upon SRH

For a diagnosis of definite diverticular hemorrhage, our patients have a SRH in a colon diverticulum on urgent colonoscopy, such as active bleeding, non-bleeding visible vessel (NBVV), an adherent clot (e.g. resistant to washing off), or a flat spot. TIC SRH are illustrated in figure 1. Some other patients classified as definitive diverticular hemorrhage have extravasation of contrast on angiography (e.g. CT angiogram - CTA, standard interventional radiology -IR – angiogram, or MRI angiogram), or extravasation of technetium labelled red blood cells (RBC) on an RBC scan while finding colon diverticula and no other lesions that could cause bleeding in that colon segment on colonoscopy performed during the same LGI hemorrhage (2). A diagnosis of presumptive diverticular hemorrhage is made if non-bleeding colon diverticula without SRH are found on urgent colonoscopy and other GI procedures are negative for a definitive lesion diagnosis in the colon, anorectum, foregut, or small intestine by anoscopy, panendoscopy, push enteroscopy, and capsule endoscopy, performed during the same hospitalization for diagnosis of a bleeding site. Incidental diverticulosis is the diagnosis if colon diverticula without SRH are seen on colonoscopy and another definitive GI source is found in the colon, anorectum, foregut, small bowel (2).

Figure 1:

Stigmata of recent hemorrhage and their prevalence rates as found in 160 patients with severe diverticular hemorrhage. These identify the diverticular bleeding site and are all criteria for diagnosis of a definitive diverticular bleed.

Patient Resuscitation, Colon Preparation, and Timing of Colonoscopy

Patients should be hemodynamically resuscitated and medically stable before urgent colonoscopy is performed (1, 2). After a rapid bowel preparation with a polyethylene glycol (PEG) based balanced electrolyte solution (1, 23), urgent colonoscopy is feasible, safe, and often diagnostic for patients with an acute, severe LGI bleed, including those with suspected diverticular hemorrhage (24). For severe hematochezia in hospitalized patients, 6 to 8 liters of a PEG solution are recommended for the patient to drink or to be given via a nasogastric (NG) tube as a liter every 30 – 45 minutes to clear the colon of stool, blood, clots, and food fiber prior to urgent colonoscopy (18, 24). To facilitate gastric emptying and reduce nausea, Metoclopramide (5-10 mg intravenously – IV) 1-2 hours before starting the colon purge and 4-6 hours later is also recommended (2, 24).

To optimize finding SRH in a diverticulum or on other colon lesions during hospitalization for an acute, severe, LGI bleed, urgent colonoscopy for diagnosis and treatment is recommended within 12-24 hours after the patient is hospitalized and evaluated by a GI bleeding team (2,5,24). Depending upon the hemodynamic stability of the patient, treatment status of active co-morbidities, and availability of anesthesia support, the colonoscopy can be performed in an operating room (OR), an intensive care unit (ICU), in a monitored bed, or in the medical procedure unit (MPU). Best results of urgent colonoscopy in finding a lesion with SRH will be when the rectal effluent is clear of blood, clots, stool, and food fiber (2,5,24).

GI Bleed Team, Equipment and Accessories

Invaluable members of a GI bleeding team are an experienced and well-trained staff including a GI endoscopy nurse, an endoscopy technician, and an anesthesiologist or nurse anesthetist. The team should be led by a gastroenterologist who is a well-trained clinician with expertise in urgent colonoscopy, GI bleeding, and colonoscopic hemostasis. All of these will contribute to successful results of urgent colonoscopies and other emergency GI procedures (25).

The choice of a standard size or smaller pediatric colonoscope depends upon the size of the patient, the completeness of colon purge, and whether colon stenosis is suspected. Current video colonoscopies have relatively large suction ports (e.g. 3.2-3.8 mm), adequate fields of view, and good tip angulation. These instruments also have a separate channel for target irrigation. Placing a cap on the colonoscope tip facilitates visualization, suctioning, and eversion of diverticula which are all useful during urgent colonoscopy for identification of SRH and diagnosis of a definitive TIC hemorrhage.

Various accessories are helpful for bedside utilization during urgent colonoscopy. These include a needle to pre-inject active bleeding or an adherent clot with epinephrine and a rotatable snare to cold guillotine the clot off prior to thermal coagulation or hemoclipping of the pedicle or underlying SRH. A colon length DEP probe with control unit is recommended for interrogation of non-bleeding TIC SRH before endoscopic hemostasis (5,18). The DEP is also useful after hemostasis of both active bleeding TIC’s and non-bleeding SRH to confirm obliteration of underlying arterial blood flow, which correlates with definitive TIC hemostasis and reduced TIC rebleeding rates (5,18). Through-the-scope hemoclips, both small (8 mm, when open) or larger (11-12 mm) are useful for focal hemostasis especially for SRH in the base of the diverticulum (5,18). Multipolar thermal coagulation (MPEC) probes in 7 and 10 French sizes are recommended with a bipolar generator for treatment of SRH at the neck of the TIC.

Over-the-scope-clips (OTSC) of various sizes, tip lengths, and different teeth configurations are also useful for hemostasis of definitive TIC bleeding and other focal colon lesions, as they are for non-variceal UGI hemorrhage (26). OTSCs grasp more tissue than through-the-scope hemoclips and this results in high rates of definitive hemostasis of bleeding lesions by obliterating arterial blood flow underneath SRH of focal lesions (26).

Injection needles and a tattoo solution to label 3-4 areas around the TIC with the SRH will facilitate identification of the bleeding site for rebleeding or surveillance colonoscopy. Multi-shot rubber band ligation (RBL) kits which fit on a colonoscope are also recommended by some colonoscopists (10-12). Other accessories which are useful in removing large clots from diverticula are a foreign body grasper and a Roth net.

Hemostasis Technique Application for Different TIC SRH

For active TIC hemorrhage, after target irrigation and suctioning of blood and clots from the bleeding site, pre-injection with dilute epinephrine (1:20,000 in saline) in 3-4 quadrants around the bleeding point in 1-2 cc aliquots is useful to temporarily stop the bleeding. If the bleeding point is on the neck of the diverticulum, MPEC for applying moderate, lateral pressure on the active bleeder to tamponade and to coagulate it is recommended, with a low power setting (e.g. 12-14 watts) with either a 7 or 10 French size probe. Coagulation with short pulses (2-4 sec) is recommended for coaptive coagulation of the underlying artery (27). Refer to figures 2- 4.

Figure 2:

Active bleeding from the neck of a colon diverticulum

Figure 4:

Firm tamponade pressure on the bleeding site and compression laterally halts arterial blood flows in the artery at the neck of the diverticulum and facilitates thermal coaptive coagulation and definitive hemostasis

If the bleeding site is in the base of the TIC, pre-injection of epinephrine followed by deployment of hemoclips, OTSC, or rubber band ligation (RBL) is recommended. Refer to figure 5 and 6. With a cap on the colonoscope or with the OTSC or RBL, TICs with the SRH in the base can be everted so that the hemoclips or band can be placed on the SRH and/or on either side of it to obliterate underlying arterial blood flow (5,10-12,24). Because the OTSCs stay on the TIC longer than through-the-scope clips or rubber bands, these large clips usually result in long term definitive hemostasis when deployed on the SRH (5,15,16).

Figure 5:

Diagram of colon diverticulum with a non-bleeding visible vessel (NBVV) in the base. The arterial anatomy is shown including a submucosal artery (*) and an interconnected subserosal artery (**)

Figure 6:

Successful colonoscopic hemostasis with hemoclip placement on either side of the non-bleeding visible vessel in the base of the diverticulum to obliterate blood flow between the clips and to prevent rebleeding

If either OTSC or RBL treatment of TIC SRH is used, initial colonoscopic diagnosis without these devices on the tip is required to first localize SRH and mark this TIC (10-12,15,16). Then a colonoscope with the OTSC or bander on the tip is reinserted to treat the SRH. For active bleeding, pre-injection with epinephrine and placement of a clip adjacent to the TIC with the SRH will help the endoscopist locate it again. Pre-loading of the OTSC or bander on the tip of another colonoscope will hasten the process of diagnostic colonoscope removal and re-intubation of a colonoscope with the OTSC or RBL attached.

For an adherent clot either in the base or neck of the diverticulum, DEP interrogation and pre-injection with epinephrine are recommended prior to shaving the clot down with a rotatable snare to a small pedicle or to expose an underlying NBVV or spot (5). If the clot is in the base of the TIC, standard hemoclipping, OTSC, or RBL are recommended but not MPEC. If the clot is at the neck, MPEC or hemoclipping are safe and effective (5). RBL or OTSC may result in definitive hemostasis more often than through-the-scope- hemoclips because these usually fall off (12,15,16).

NBVVs and flat spots at the neck or in the base of the TIC are treated to similar to adherent clots (5,20). However, for these SRH, pre-injection with epinephrine is not necessary unless bleeding is induced. DEP interrogation of the SRH before treatment of non-bleeding SRH and after treatment of both non-bleeding and active bleeding is recommended (5,18). If there is residual arterial blood flow detected by DEP after visually guided hemostasis, either additional MPEC treatment or hemoclipping to obliterate the arterial blood flow underneath SRH will result in definitive hemostasis and will reduce the risk of rebleeding (5,18).

Tattooing the adjacent mucosa around the TIC with the SRH in 4 quadrants after colon hemostasis is recommended. This will facilitate identifying the TIC in case of early rebleeding or during surveillance colonoscopy.

Medical Treatment

Medical treatment to control constipation with daily supplemental fiber, a stool softener, and drinking an extra liter of fluid (without alcohol or caffeine) is recommended (2,5,18). Anti-thrombotic drugs (e.g. aspirin or Clopidogrel), anti-coagulants, and non-steroidal anti-inflammatory drugs (NSAID’s) should be stopped in the short term and longer if the prescribing subspecialist (e.g. cardiologist, neurologist, or vascular surgeon) agrees that this is safe (5,18,28).

RESULTS

Non-Colonoscopy Tests and Results

Although RBC scanning has been reported to detect active bleeding at a very low rate of 0.04 cc/min (29), this test is not readily available as an emergency procedure in most smaller hospitals and there are other limitations for diagnosis of diverticular hemorrhage. First, an anatomic diagnosis is not possible without other imaging. Second, because only about 30% of patients with definitive diverticular hemorrhage have active bleeding on urgent colonoscopy and because RBC scanning and other radionuclide tests do not detect nonbleeding SRH, 70% or more of the RBC scans will be negative and not detect a bleeding site, including a definitive diverticular hemorrhage (2,5,18). A positive RBC scan within 30-60 minutes after radionuclide injection is helpful as a screening test prior to angiography to potentially increase the yield of angiography (1,2,24). However, scans that are positive later (e.g. 4-24 hours) after injection of the radionuclide may be inaccurate for lesion localization because of movement of the label in the colon with peristalsis in between scheduled scans (30). Although the diagnostic yield for patients with different types of GI hemorrhage is reported to be as high as 45% (29,30), CURE Hemostasis studies of TIC hemorrhage report that the RBC scan yield is less than 20% (1, 2, 5,24,25).

CT angiograms (CTAs) are usually available in larger referral hospitals and are often ordered from the emergency room (ER) for patients presenting with acute severe LGI hemorrhage. Although there are reports of diagnostic yields of 50% or higher (31,32), CTAs have many of the same limitations as RBC scans. First, only active bleeding can be detected by CTA and not non-bleeding SRH in TIC’s and on other LGI lesions (1,2,5,18,33). Second, CTA is not recommended in patients with renal insufficiency because acute worsening of renal function may result. Third, CTAs which are not therapeutic expose the patient to radiation when other tests such as colonoscopy or angiography may be more useful clinically because they can be both diagnostic and therapeutic for TIC hemorrhage.

Angiography has theoretical advantages of detecting bleeding when active bleeding is 0.5 cc/minute or more (34). When extravasation of contrast is seen, embolization of the focal bleeding site is feasible (34-36). Disadvantages and limitations are lack of availability of urgent angiography in smaller hospitals and a low pre-test probability of a positive test since only about 30% of definitive TIC hemorrhage have active bleeding on urgent colonoscopy (See Figure 1). Also, angiography is expensive and major complications can occur. Complications include acute kidney injury, bowel ischemia, hematomas, artery thrombosis, and allergic reactions to contrast agents (36). As an example, in one study where acute LGI bleeding was controlled by emergency angiographic embolization, severe gut ischemia and death were reported for several patients (36). Also, angiography may not be clinically useful because it is only diagnostic for a bleeding site in less than 25% of patients with severe LGI bleeds (1) and less than 20% of patients with colonoscopically documented definitive diverticular hemorrhage (2). However, angiography with embolization is particularly useful as salvage treatment after a definitive TIC hemorrhage is treated with hemoclips and early rebleeding (e.g. within 30 days) occurs. Even if active bleeding is not seen on the angiogram, the hemoclips can serve as a target for super selective embolization of the TIC with the rebleeding (18, 37).

Surgery

With current colonoscopic, radionuclide scanning, and angiographic techniques, surgery is now rarely required as a first line for diagnosis and treatment of TIC hemorrhage (1,2,5,18,37). It is indicated as a salvage treatment for TIC hemorrhage which is ongoing or recurrent and when colonoscopic hemostasis and angiography with embolization are unsuccessful (18,37). Surgery is also indicated for severe colon ischemia after IR embolization or for a perforation after colonoscopic hemostasis. Permanent hemostasis can be expected after resection of the bleeding TIC segment, but the complication rate of urgent surgery is significantly higher than IR embolization or urgent colonoscopy in a recent report (37).

Colonoscopy

Diagnostic Yield

Colonoscopy has the advantages of being widely available, safe, and often diagnostic in patients hospitalized with acute LGI bleeding (1,2,5,18,24,38, 39). Colon hemostasis is also safe and effective when a definitive TIC bleed is diagnosed on urgent colonoscopy based upon SRH and supplemented by DEP monitoring of underlying arterial blood flow (5,18). In an initial report of urgent colonoscopy for acute severe LGI hemorrhage, the diagnostic yield was 80% compared to a 20% yield by angiography when these tests were both performed in the same patient (1). Refer to figure 1 which documents the SRH findings on urgent colonoscopic for 160 patients with definitive TIC hemorrhage. An adherent clot was found in 40% of patients, active bleeding in 30%, a NBVV in 20% and a flat spot in 10%.

When the DEP probe became available for monitoring arterial blood flow in the colon, TIC SRH were evaluated to potentially improve risk stratification and to increase the rate that definitive diverticular hemorrhage was diagnosed during urgent colonoscopy (18,20). Recent results are shown in table 2 that include flat spots as a major SRH. Detection rates of arterial blood flow were very high underneath TIC SRH, varying from 80% for flat spots or adherent clots to 93% for NBVV and 100% for active bleeding. The overall detection rate of arterial blood flow for these SRH was 88%, which contrasted with a 0% detection rate in presumptive TIC bleeds when the base of the TICs with small visible arteries without SRH were interrogated (18).

Table 2. Arterial Blood Flow by Doppler Probe for Definitive vs. Presumptive Diverticular Hemorrhage.

*p < 0.05. + DEP is positive Doppler endoscopic probe signal for arterial blood flow. SRH are stigmata of recent hemorrhage. NBVV is non-bleeding. TIC is diverticular (bleed). See references (18, 20)

Stigmata of Hemorrhage	Number	+ DEP	Totals (Rate + DEP)
Major SRH
Active Bleeding	5	5	5/5 (100 %)
NBVV	14	13	13/14 (93 %)
Clot	5	4	4/5 (80 %)
Flat Spot	10	8	8/10 (80 %)
Definitive TIC Bleed	34	30	30/34 (88 %)*
Clean Base
Presumptive TIC Bleed	32	0	0/32 (0% )

In a report from 2016 of 436 patients hospitalized with colon diverticulosis and severe hematochezia, the distribution by final TIC diagnosis is illustrated in Figure 7. Incidental diverticulosis (e.g. a non-diverticular source was the cause of the LGI bleed) was the final diagnosis in 45.6%. Presumptive diverticular bleeding was the diagnosis in 28%. Definitive diverticular TIC bleed was diagnosed in 26.4% of patients based upon finding a SRH in a colon diverticulum. Subsequently, the CURE Hemostasis Research group updated their results for the diagnosis of diverticular hemorrhage as shown in figure 8. For 275 patients hospitalized with severe LGI bleeding and colon diverticulosis who were evaluated with urgent colonoscopy, SRH, and DEP, 75% had presumptive TIC hemorrhage and 25% had definitive TIC hemorrhage.

Figure 7:

Prevalence of definitive, presumptive, and incidental diverticular hemorrhage. 436 patients hospitalized with diverticulosis and severe hematochezia.

Figure 8:

Proven diverticular hemorrhage: recent prevalences of definitive and presumptive bleeding based on urgent colonoscopy. 100 consecutive patients with severe hematochezia were evaluated by urgent colonoscopy by the CURE hemostasis Research Group.

Definitive TIC Hemorrhage: Natural History and Colonoscopic Hemostasis Results

In a 2016 report about the natural history of definitive diverticular hemorrhage in patients treated medically without colon hemostasis, results are summarized in table 3. For patients with definitive diverticular bleeds, the rates of more, severe TIC bleeding within 30 days varied by SRH. That is from 43% for adherent clot patients to 84% for those with active TIC bleeding. Rates of intervention within 30 days to control TIC bleeding also varied by SRH from 29% for adherent clot patients to 58% for those with actively bleeding TICs (18,20). Based upon these high rates of rebleeding on medical therapy and high rates of major intervention (e.g. with angiography or surgery in most patients or more recently repeat urgent colonoscopy by an expert colonoscopist), colonoscopic hemostasis is highly recommended when any of these SRH are found during urgent (or later) colonoscopy.

Table 3. Natural History and 30 Day Outcomes of Definitive Diverticular Hemorrhage on Medical Therapy.

N is number of patients. NBVV is non-bleeding visible vessel. Clot is adherent clot. Interventions were either angiography with embolization, surgery, or repeat urgent colonoscopy and hemostasis by an expert colonoscopist.

	Major Rebleed	Intervention For Rebleed
Active bleed (N = 19)	84%	58%
NBVV (N = 5)	60%	40%
Clot (N = 14)	43%	29%
Flat spot (N =6)	60%	33%
TOTALS (N = 44)	64%	43%

In a 2018 report of 118 patients with definitive diverticular hemorrhage, early rebleeding rates were compared for patients treated with colonoscopic hemostasis or medical therapy alone (5). The definitive TIC treatment groups included patients with active bleeding, NBVV or adherent clots, but not flat spots (5). For 81 patients with colonoscopic hemostasis, the 30-day rebleeding rate was 6.2%, whereas 37 patients treated medically had a rebleeding rate of 64.9% (5). For the medical and colonoscopic hemostasis groups, the rate of surgery or embolization for control of severe rebleeding within 30 days were 43.2% and 2.5% respectively. The median days to discharge after treatment were also significantly higher in the medical group than the colon hemostasis group (8.5 vs 2 days).

In a recent report, 162 patients with definitive TIC hemorrhage and 3 or more months of follow-up had clinical outcomes compared according to medical treatment alone, embolization angiographic, surgery, or colon hemostasis (37). The rebleeding rates within 30 days of angiographic embolization, surgery, colonoscopic hemostasis, or medical treatment alone were 6.7%, 0%, 8.7%, and 41.7%. The respective complication rates were 13.3%, 36.9%, 2.9% and 0%. The median hospital days were 6 days, 12 days, 2 days and 4 days respectively. During one year of follow-up, the rates of rebleeding were 0% for surgery, 13.3% for IR embolization, 17.3% for colonic hemostasis, and 50% for medical treatment. Based upon these studies, urgent colonoscopy was strongly recommended for early diagnosis and hemostasis of definitive TIC hemorrhage. These recommendations are similar to those of the Japan Gastroenterological Association (38). In the CURE TIC study, either DEP monitored treatment or OTSC appeared to yield higher long-term rates of definitive TIC hemorrhage (37). However, no RCTs have been reported which compare these newer treatments with other older TIC treatments.

Presumptive TIC hemorrhage

The CURE Hemostasis Group recently reported on the natural history of 158 patients with a colonoscopic diagnosis of presumptive TIC hemorrhage (39). During a median follow-up of 59 months. 71.5% of patients had no TIC rebleeds and 28.5% had 1 or more TIC rebleeds. For rebleeds, 40% were diagnosed as definitive and 60% were presumptive again. Definitive TIC hemorrhage diagnosis was made by urgent colonoscopy in 44.4%, RBC scanning in 27.6%, angiography in 16.2%, and surgery in 11.1% (39).

DISCUSSION

Most patients with diverticular hemorrhage are elderly with the mean age of 74 years (1,2,5,18,24). Not all patients who present to the hospital with severe hematochezia have colonoscopy within 6-24 hours when TIC SRH are the most prevalent (1,2,5,18). However, when urgent colonoscopy is performed by experienced colonoscopists such as in the US or Japan, the rate of diagnosing definitive TIC hemorrhage is approximately 25% (18,23,38). Even though the rate of diagnosing definitive TIC hemorrhage is higher during long term FU of patients with a prior presumptive TIC hemorrhage at 40% (39), this is much lower than the definitive diagnosis rate of urgent pan endoscopy for severe UGI hemorrhage (40).

Definitive diagnosis during colonoscopy of TIC hemorrhage depends upon the quality of the colon prep, the timing of the colonoscopy after the LGI bleed, and most of all the expertise of the colonoscopist in recognizing SRH (5,18,25). When there is a dedicated GI bleeding team prepared to perform urgent colonoscopy, higher rates of diagnosis and colonoscopic hemostasis are more likely (2,5,18,25,38). However, without other major advances to improve early diagnosis of definitive TIC hemorrhage, it is unlikely that the rate of definitive TIC hemorrhage diagnosis will exceed 40%.

There is a lack of consensus about the yield of urgent colonoscopy for diagnosis of acute LGI bleeding and whether the timing makes a difference in rates of finding SRH on diverticula (25,38,41-43). In a retrospective report from the Mayo Clinic in 2003, there was no difference in the rate of finding TIC SRH (e.g. active bleeding, NBVV, or adherent clot), for different colonoscopy times – either early, later, or elective (41). In an RCT from 2010, Laine et al reported that there was no difference in outcomes (e.g. more bleeding, transfusions, hospital stay, or hospital changes) for patients with acute LGI bleeding who had urgent colonoscopy within 12 hours, or elective colonoscopy more than 36 hours after hospitalization (42). Limitations of this RCT are that most patients did not have diverticular hemorrhage and those with severe LGI hemorrhage were excluded from enrollment (42). In a recent guideline from the American College of Gastroenterology (ACG), urgent colonoscopy was not recommended for patients with acute LGI hemorrhage (43). That contrasts with the recommendations of the CURE Hemostasis Research Group (1,2,5,18), some reports by experts from Japan (10-12), and the Japan GI Association (38).

In another report of acute LGI bleeding, early colonoscopy resulted in shorter hospitalization (44). In CURE prospective studies of acute LGI hemorrhage, the hospitalization time and cost of urgent colonoscopy cases were less than elective colonoscopy or other treatments (6). Also, in a more recent report of patients with definitive diverticular hemorrhage, the length of hospitalization with colonoscopic hemostasis was significantly shorter than for those patients treated medically (5).

As future directions, there are three recommendations which are complimentary. The first is to amplify the didactic and procedural training in LGI bleeding (25). Because of earlier reports (41,42) and a recent ACG guideline on LGI bleeding (43), urgent colonoscopy is not currently recommended for diagnosis by some gastroenterologists or the ACG. This will contribute to limiting the training about acute LGI hemorrhage, deemphasizing discussion about this topic in GI meetings, and ultimately reducing the utilization rate of urgent colonoscopy for diagnosis of definitive diverticular hemorrhage in the US. In addition, most hospitals do not support funding of a GI bleeding team which can improve patient care, clinical outcomes, promote teaching of GI trainees, and reduce patient care costs (25). Such funding should be available, particularly for large referral centers.

Second, there is a need for innovations for earlier diagnosis of definitive TIC and other colon bleeding sites by detecting either active bleeding and other non-bleeding SRH before these disappear. One candidate is urgent colon capsule endoscopy which can be performed when a patient presents to the ER with a severe acute LGI bleeding. In initial results by the CURE hemostasis group, more than 80% of patients had a lesion diagnosis and/or bleeding site localization by urgent colon capsule compared to less than 25% for CTA or RBC scanning (45). Colonoscopy or other tests confirmed the final diagnosis which included TIC hemorrhage and other colorectal lesions. The advantages of colon capsule for urgent diagnosis and localization of the bleeding site are multiple. First, colon capsule endoscopy is ingested without sedation and the procedure can be performed in either small or larger hospitals. After downloading images, telemedicine can be utilized for remote image review and diagnosis by experts. Second, this can be performed early while waiting for other procedures to be scheduled. Also, the colon capsule results can be utilized as a management guide for decisions about ordering other GI endoscopic procedures, IR embolization, or surgery. Last, urgent colon capsule endoscopy is relatively inexpensive compared to CTA, standard angiography, or colonoscopy and is less labor intensive. Future studies are warranted to determine what the role of urgent colon capsule endoscopy may be and whether it can increase the diagnostic yield of definitive diverticular hemorrhage and other colon sites.

Last, the role of elective colon surgery with laparoscopic resection of segmental diverticulosis for patients with recurrent diverticular hemorrhage should be evaluated. In the CURE Hemostasis Group’s experience, there are several relevant results and reasons to recommend this. First, most patients with definitive diverticular hemorrhage have the TIC bleeding site located at or proximal to the splenic flexure. However, most colon diverticula are anatomically distributed distal to the splenic flexure in most US patients (2,5,18). Second, most early or recurrent diverticular rebleeds are from the same colon TIC and not another region of the colon (5,24). Third, elective laparoscopic segmental colectomy is much safer than emergency colon resection and should be as effective in preventing long term TIC rebleeding (1,2,24). Last, in patients having colon surgery for a definitive diverticular hemorrhage, the CURE Hemostasis Group reported no TIC rebleeding during long term follow up (37). For long term prevention of TIC rebleeding, surgery is more definitive treatment than colonoscopic hemostasis or IR embolization (37). For patients with recurrent TIC hemorrhage who are surgical candidates, a new algorithm should be evaluated for elective laparoscopic resection of the colon segment where the bleeding site was localized preoperatively. That is particularly relevant in patients with risk factors for TIC rebleeding, including those on chronic NSAID’s, aspirin, other anti-platelet drugs, or anti-coagulants risk (46,47).

CONCLUSIONS

There are multiple conclusions of this analysis about the diagnosis and treatment of diverticular hemorrhage. First, when urgent colonoscopy after adequate colon purge is performed by an expert colonoscopist, the diagnostic rate of definitive diverticular hemorrhage will be about 25% and for patients with a prior presumptive diverticular hemorrhage who rebleed will be about 40% (2,39). Second, when definitive diverticulosis is based upon SRH, the most common findings will be adherent clot (about 40%), active bleeding (30%), NBVV (20%), and a flat spot (10%). Third, DEP supplements SRH for risk stratification before hemostasis and for documenting definitive hemostasis with obliteration of arterial blood flow underneath TIC SRH (5,18). Fourth, when colonoscopic hemostasis is successful, shorter hospitalization and lower complications are reported when compared to medical treatment, IR embolization or surgery (2,5,6,18,44). Fifth, for failures of colonoscopic diagnosis and hemostasis, angiographic embolization or colonic surgery are recommended (37). Last, improvements in early diagnosis of definitive TIC hemorrhage are warranted and so is evaluation of an algorithm to incorporate elective surgery with segmental resection in patients with recurrent TIC hemorrhage when the bleeding TIC or colon segment has been localized pre-operatively.

Figure 3:

Pre-injection of a dilute epinephrine around the bleeding site will stop active bleeding prior to use of the MPEC probe for definitive hemostasis

Synopsis:

This is a description and critical analysis of current diagnosis and treatment of diverticular hemorrhage. The focus is on colonoscopy for identification and treatment of stigmata of recent hemorrhage (SRH) in diverticula. A classification of definitive, presumptive, and incidental diverticular hemorrhage is reviewed and recommended. The approach to definitive diagnosis with urgent colonoscopy is put into perspective of other management strategies including angiography (of different types), nuclear medicine scans, surgery, and medical treatment. Advancements in diagnosis, risk stratification and colonoscopic hemostasis are described including those that obliterate arterial blood flow underneath SRH and prevent diverticular rebleeding. Recent innovations are discussed.

Key Points:

1) Colonoscopy is the most accurate method to diagnose and treat diverticula hemorrhage. 2) Risk stratification with diverticular stigmata of hemorrhage and monitoring underlying arterial blood flow facilitate accurate diagnosis and definitive colon hemostasis. 3) Clinical outcomes are improved by urgent colonoscopic diagnosis and treatment of definitive diverticular hemorrhage.

Clinics Care Points.

• With urgent colonoscopy in well prepped patients with colon diverticulosis and severe LGI bleeding, experienced colonoscopists can expect to find TIC stigmata in about 25% of cases, for a diagnosis of definitive TIC hemorrhage. If no TIC stigmata or other lesions are found, presumptive TIC hemorrhage will be the diagnosis in about 75% of patients with true diverticular hemorrhage,

• The expected prevalence of TIC stigmata in definitive TIC hemorrhage cases is adherent clots (40%), active bleeding (30%), visible vessels (20%) and flat spots (10%).

• Don’t expect any type of angiography (CTA, MRI, or standard types) or RBC scan to be able to diagnose more than the 30% of definitive TIC hemorrhages because non-bleeding stigmata are not identified by these imaging techniques.

• Expect to find about 50% of TIC stigmata in the diverticulum base and the other 50% at the neck of the TIC. You may have to use a cap to evert the TIC to see and treat SRH in the base.

• For efficacy and safety, use hemoclips, OTSC, or rubber band ligation to treat stigmata in the TIC base and MPEC or OTSC at the neck. If you confirm absence of arterial blood flow under the SRH after hemostasis, the TIC rebleed rate will be very low unless the hemoclips or rubber bands fall off early.

• Embolization and surgery are effective as salvage therapies for severe rebleeding.

ABBREVIATIONS

ACG

American College of Gastroenterology

CT

computerized tomography

CTA

CT angiogram

DEP

Doppler endoscopic probe

ER

emergency room

FU

follow-up

GI

gastrointestinal

ICD

International Classification of Diseases

ICU

intensive care unit

IR

Interventional Radiology

IV

intravenous

LGI

lower gastrointestinal

MPU

Medical Procedure Unit

MRI

Magnetic Resonance Imaging

NBVV

non-bleeding visible vessel

NG

nasogastric (tube)

NSAIDS

non-steroidal anti-inflammatory drugs

OR

operating room

OTSC

over-the-scope-clip

PEG

polyethylene glycol (colon preparation solution)

PUB

peptic ulcer bleeding

RBC

red blood cell

RBL

rubber band ligation

SRH

stigmata of recent hemorrhage

TIC

diverticular