Keywords

Mechanism of NSAIDs

Most NSAIDs are nonselective inhibitors of cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2) enzymes. While aspirin acetylates COX-1 and COX-2 irreversibly, NSAIDs do so reversibly. COX-1 is constitutively expressed and produces mucosal protective prostaglandins to the gastric and duodenal mucosa. In addition, COX-1 contributes to platelet aggregation through thromboxane production. COX-2 is inducible and plays a lesser role in protection of mucosal lining from acid and pepsin. On that basis, it has been thought that COX-2 selective inhibitors have a safer GI side effect profile. However, at clinically significant doses, COX-2 inhibition leads to similar adverse gastrointestinal effects. Moreover, animal studies have refuted this “COX-2 hypothesis”, suggesting that both COX-1 and COX-2 inhibition are required for ulcer formation. [8, 21, 28] Therefore, it is not the selective COX-2 inhibition alone, but perhaps the absence of dual inhibition of both isoforms that explains the decreased GI side effects with COX-2 selective NSAIDs such as celecoxib.

Most NSAIDs are completely absorbed and have trivial first pass hepatic metabolism. Side effects including renal, cardiovascular and gastrointestinal vary, and are largely dependent on pharmacodynamics and pharmacokinetics [27]. It is also important to note that NSAIDs notoriously interact with commonly used medications such as methotrexate and ACE inhibitors. Figures 15.1 and 15.2 illustrate the inhibition of COX-1 and COX-2 and the downstream effects.

Fig. 15.1
figure 1

Mechanism of NSAIDs. https://springerlink.bibliotecabuap.elogim.com/article/10.1007/s40629-018-0064-0

Full size image
Fig. 15.2
figure 2

Mechanism of NSAIDs. https://media.springernature.com/original/springer-static/image/art%3A10.1007%2Fs40266-019-00660-1/MediaObjects/40266_2019_660_Fig 1_HTML.png

Full size image

Pathogenesis of NSAID-Induced Enteropathy and Colopathy

While some studies report 15–30% of patients on chronic NSAIDs develop peptic ulcer disease, other studies observed as high as 80% of long term users eventually develop NSAID-induced enteropathy [2, 6].

Aspirin has a pKa of 3.5 and ibuprofen 4.85. The more positive the pKa value, the weaker the acid and so therefore, aspirin and NSAIDs are not ionized in the gastric lumen but rather can travel across the mucosa. As the mucosa has a neutral pH, NSAIDs then linger in the epithelial cells and can damage the epithelial cells through this mechanism. The second and most critical means of destruction is by inhibiting COX-1 and COX-2, thereby decreasing prostaglandin synthesis. Aspirin dosages as low as 10 mg/day can inhibit gastric prostaglandins [8, 24]. In turn, human gastric mucosa can require up to 8 days to restore its COX-1 activity. Without prostaglandins, the mucosal layer loses its cytoprotective properties and is now susceptible to unimpeded gastric acid secretion, proteolytic enzymes, bile acid and toxins [6]. First, prostaglandins stimulate epithelial cell secretion of mucin and bicarbonate which together form a protective alkaline barrier. Secondly, prostaglandins also enhance mucosal blood flow and oxygen through vasodilation. In addition, prostaglandins promote epithelial cell proliferation and movement towards the luminal side [8]. Other proposed protective mechanisms include interference with nitric oxide such that some studies are now investigating NSAIDs coupled with nitric oxide and aspirin coupled with nitric oxide. In addition to interfering with a variety of gastroduodenal defensive pathways, NSAIDs also promote subsequent bleeding from peptic ulcer disease by inhibiting platelet aggregation and vasoconstriction.

With increased utilization of capsule endoscopy, small bowel enteropathy is more and more frequently identified. Pathogenesis in contrast to gastroduodenal disease, involves intestinal gram-negative “dysbiosis” and bile toxicity as seen in Fig. 15.3, particularly in those with NSAIDs plus a proton pump inhibitor (PPI) [2, 8]. In animal studies, NSAIDs with PPI causes disruption of the microbiome, with overgrowth of gram-negative bacteria in the ileum and subsequent ulceration while those with gram-positive bacteria such as Bifidobacteria and Firmicutes tend to be without ulcers [8]. Similarly, human studies such as Washio et al. study showed that COX-2 selective NSAIDs in combination with PPI had significantly higher incidence of ulcers, particularly in the jejunum as compared to COX-2 selective NSAIDs and placebo due to small bowel dysbiosis [8]. Interestingly, current prospective studies including GI-REASONS trial and CONDOR trial are also finding that occult GIB is more common in nonselective NSAIDs coupled with PPI compared to celecoxib alone [8].

Fig. 15.3
figure 3

Pathogenesis of NSAID-induced enteropathy. https://springerlink.bibliotecabuap.elogim.com/article/10.1007/s10620-015-3963-7

Full size image

In the lower GI tract, NSAIDs, especially enteric-coated and slow release formulations which permit more exposure of the drug in the colon, increase colonic ulcers, diverticular bleeding and exacerbations of inflammatory bowel disease [12]. In comparison to gastric and small bowel ulcers, colonic ulcers are much less common, less described and found in less than 5% of patients on chronic NSAIDs [9]. There is also a predilection for right-sided and proximal colon. With increasing use of the sustained release formulation, more drug reaches the cecum which then acts as a “reservoir” and becomes directly toxic to the right colon [11, 12, 14, 15]. Histologic exams have reported pill-coated NSAID particles in histiocytes in ulcer and granulation tissue biopsied in the right colon. Distally, proctitis can develop with rectal NSAIDs. Primary mechanisms for NSAID-induced colopathy are decreased blood flow and disruption of epithelial cells as well as decreased prostaglandin production by colonic mucosa [15]. Disruption of epithelial cells leads to exposure to and bacterial translocation and inflammatory reactions. Colopathy can occur within days of NSAID use [11].

Although much of literature focuses on oral NSAIDs, it is important to remember that NSAIDs also come in topical formulations. There have been case reports where topical NSAIDs have also led to occult GI bleeding. Hirose et al. reported a case of an occult GI bleed in an elderly woman using eight sheets of 20 mg of ketoprofen patches every day [5].

Risk Factors for NSAID-Induced Enteropathy

Dosage , duration and individual risk factors and co-morbidities dictate the risk of developing peptic ulcers. Expectedly, both increased dosage and duration augment the risk. An overall average of three months is generally noted but even continuous use during a two-week period can cause mucosal injury. Oral NSAIDs in conjunction with glucocorticoids, warfarin, antiplatelets, topical NSAIDs, and SSRIs potentiate GI bleeding. The following factors increase risk of GI toxicity: previous ulcer disease, age greater than 60, high dose NSAIDs combined with steroids, antiplatelet agents and anticoagulants, chronic NSAID use, untreated H. pylori infection, hemodialysis and SSRIs [8]. In 2009, the ACG stratified risk of NSAID-induced PUD into low, moderate and high as seen in Fig. 15.4 below.

Fig. 15.4
figure 4

Risk stratification of gastroduodenal toxicity in NSAID use. (Adapted from 2009 American College of Gastroenterology)

Full size image

Clinical Presentation and Complications

Melena and hematemesis from NSAID-induced ulcers are well known consequences. Close to 15% of peptic ulcers bleed with a mortality rate of roughly 10%. Patients with peptic ulcers can experience epigastric pain or remain asymptomatic, with many eventually presenting with occult GI bleeding or iron deficiency over time [3]. 51Chromium-labelled erythrocyte studies in NSAID-induced enteropathy cases estimate occult blood loss from one cc to up to 10 cc a day [1]. When symptomatic, patients usually describe the pain as a hunger sensation with pain worse after eating in gastric ulcers and better with eating in the case of duodenal ulcers. The emergence of early satiety and vomiting suggests obstruction secondary to the ulcer itself or formation of diaphragms. Diaphragms often develop in the ileum but can occur throughout the small bowel, as seen in Fig. 15.5 [6]. Dysphagia may be due to peptic strictures. Worsening abdominal pain may be the initial clue to perforation.

Fig. 15.5
figure 5

(a) Small bowel diaphragm. (b–d) Small bowel NSAID-induced ulcers. https://springerlink.bibliotecabuap.elogim.com/article/10.1007/s00535-009-0102-2

Full size image

In NSAID-induced colopathy, iron deficiency anemia, change in bowel habits, abdominal pain are reported signs and symptoms [12, 14, 25]. Like the upper GI tract, ulcers leading to occult or overt GI bleed and strictures leading to obstruction can also form in the colon.

Diagnosis

Surrogate markers such as fecal excretion of 111-Indium, fecal calprotectin, and 51Chromium-labelled erythrocyte studies can indicate intestinal inflammation due to NSAIDs [8, 24, 28]. In current practice, endoscopy is the standard for diagnosis. NSAID-induced enteropathy can range from red spots, erosions to round, annular or linear ulcers seen during upper endoscopy [1]. While varied in appearance, these ulcers are usually superficial and less than one centimeter in diameter. Most duodenal ulcers will form in the duodenal bulb and gastric ulcers in antrum or pyloric channel. In the colon, ulcers can occur throughout but rarely on the ileocecal valve.

The first diagnostic step in occult GI bleeding is upper endoscopy and/or lower endoscopy. If unrevealing, capsule endoscopy or small bowel enteroscopy should be utilized to evaluate the small bowel [1]. Studies have shown the positive predictive value of small bowel enteroscopy exceeds that of a capsule endoscopy when looking for small bowel ulcers, however more large-scale studies are still needed. Regardless, capsule endoscopy studies have estimated up to 50–80% of NSAID-induced enteropathy can occur in the small bowel. Capsule studies have seen small bowel erosions and ulcers in as little as seven days of 100 mg of enteric-coated aspirin. Jejunum and ileum are affected equally. Equally important to note, is that discontinuation of aspirin and other NSAIDs can lead to negative capsule study, especially as small erosions and superficial ulcers can heal quickly with cessation of culprit medications [7].

In the colon, NSAID-induced ulcers typically occur in the right colon and transverse and less so in the left colon [10]. They are usually well-circumscribed, circular or semicircular. Most ulcers heal within four weeks of drug withdrawal [10]. Persistent ulceration, healing and fibrosis lead to diaphragm-like strictures, especially in the right colon [13].

Prevention and Management

Perhaps the best and ideal strategy is to avoid NSAIDs in patients already at increased risk for peptic ulcer disease. These patients include those with H. pylori, Zollinger-Ellison syndrome, inflammatory bowel disease, sarcoidosis, CKD, CMV and HSV infections. However, given the myriad chronic conditions including cardiovascular and rheumatologic diseases, avoidance may be next to infeasible and so additional strategies can be incorporated in preventing and managing NSAID-induced enteropathy and colopathy.

Screening for occult bleeding in those taking NSAIDs or anticoagulants through biomarkers such as FOBT is not recommended [17]. However, testing for and eradicating H. pylori is recommended in those requiring prolonged NSAID use. Proton pump inhibitor remains the mainstay primary and secondary prevention strategy. In those who must continue on NSAIDs or aspirin for long term, PPI should be prescribed for as long as they are on those medications. There is no overwhelmingly convincing data that one PPI is superior than another PPI. As previously mentioned, while protective of the gastroduodenal tract, PPI in combination with NSAIDs can lead to small bowel ulcers through intestinal dysbiosis. This however, does not change the current recommendations in adding a PPI to long term NSAID and aspirin users.

Despite conflicting evidence, some practitioners still gravitate towards selective or relatively selective COX-2 inhibitors such as celecoxib, etodolac, diclofenac and meloxicam rather than nonselective NSAIDs in high risk patients. Although not rigorously studied, high risk patients who must take aspirin and NSAIDs together, can derive benefit from switching to selective COX-2 inhibitor plus PPI versus nonselective NSAID plus PPI. Of note however, selective COX-2 inhibitors do not provide any added benefit in decreasing lower GI bleeding related to NSAIDs [9].

In the lower GI tract, most ulcerations heal but strictures usually do not and may require balloon dilation and even surgical intervention or segmental colectomy if symptomatic [11, 15].

Alternative Therapies

While avoidance of NSAIDs and aspirin seems impossible, one alternative is utilizing topical NSAIDs in cases of pain relief, with the idea that there are less systemic effects. PPIs are preferred in prevention of peptic ulcer disease in NSAID users, but there are several other alternatives, although inferior, which can serve as adjuvant therapy. Misoprostol is a prostaglandin E analog. A study showed a relative risk reduction of 40% in bleeding and complications related to NSAIDs and aspirin when patients were taking 200 mcg four times a day for at least six months compared to placebo. A more recent 2018 randomized, double-blind, placebo-controlled trial conducted by Taha et al. looked at obscure GI bleeding in patients taking aspirin and NSAIDs and found that there was a statistically significant difference in healing of small bowel ulcers and erosions in the misoprostol group compared to the placebo group after eight weeks of therapy [4]. Unfortunately, misoprostol is not well tolerated due to substantial GI side effects including abdominal pain and diarrhea. Nonetheless, it may be worthwhile to consider misoprostol in patients with recurrent peptic ulcer disease on NSAIDs.

Sucralfate is another adjuvant therapy that can be effective in treating particularly duodenal ulcers by forming an adhesive “bandage” at the site to encourage healing. In addition, sucralfate can also impede pepsin activity in gastric secretion.

There are several other new alternatives currently being studied. Rebamipide enhances intracellular prostaglandin synthesis and has been shown to reduce NSAID-induced enteropathy in animals in the STORM trial [8]. Tranexamic acid, dipeptidyl peptidase inhibitors, geranylgeranylacetone which facilitates mucous production, irsogladine a PDE inhibitor which promotes gap junction integrity, nitric oxide releasing NSAIDs are all under investigation. Given evidence that PPI can add insult to NSAID-induced enteropathy through intestinal dysbiosis, studies such as Satoh et al. have also looked at soluble fiber and probiotics as a preventative approach [22].

Anticoagulation: Mechanism and Pathogenesis

The mechanism of warfarin, as seen in Fig. 15.6, is inhibition of vitamin K dependent clotting factors (II, VII, IX, X), protein C and protein S and thereby systemic anticoagulation. Apixiban, rivaroxaban and edoxaban are factor Xa inhibitors which prevent thrombin formation. Dabigatran inhibits thrombin, which in turn impedes fibrin production.

Fig. 15.6
figure 6

Mechanism of warfarin and DOACs. https://springerlink.bibliotecabuap.elogim.com/article/10.1007/s40800-016-0031-y

Full size image

But in addition to its action in the coagulation cascade, DOACs also have a direct effect on intestinal mucosa as most are only partially absorbed. In particular, the tartaric acid in dabigatran is erosive to the gastroduodenal mucosa [18].

Warfarin can be reversed using Vitamin K, fresh frozen plasma or prothrombin complex concentrate. Reversal agents available for dabigatran include hemodialysis and idarucizumab, which is a humanized monoclonal antibody fragment targeted against dabigatran. For the other DOACs, andexanet alfa is a recombinant modified protein mimicking factor Xa and thus binds apixiban, rivaroxiban and edoxoban. Aripazine is a cation and still being studied in reversal of all DOACs.

Fig. 15.7
figure 7

Mechanism and reversal agents of warfarin and DOACs. *Aripazine is under investigation

Full size image

Overall, DOACs are associated with decreased risk of mortality due to fatal bleeding but associated with higher risk of GI bleeding compared to warfarin [18, 19]. Compared to warfarin, dabigatran and edoxaban at higher dosages and standard dose rivaroxaban are associated with higher gastrointestinal bleeding events. Rivaroxaban is thought to be higher risk for GIB due to its higher peak serum concentrations.

Unlike NSAIDs which tend to affect the upper tract, DOACs tend to affect the lower GI tract. This is seen in the RE-LY trial looking at dabigatran, particularly given its incomplete absorption in the upper GI tract and therefore subsequent erosive effect in the lower tract [19]. In addition to its obvious effect on systemic anticoagulation, current phase 3 trials have identified additional risk factors for bleeding including age greater than 75, history of GIB, diverticulosis, angiodysplasias, renal disease, use of antiplatelets or NSAIDs and a HAS-BLED score of equal to or greater than 3 [19].

Etiology of Occult GI Bleeding in the Setting of Anticoagulation

A prospective study by Jaffin et al. suggested that those on anticoagulants with occult bleeding were not due to the anticoagulant itself but pre-existing intestinal and colonic lesions [9, 16]. Many studies such as the COMPASS trial have looked at the higher incidence of GIB with anticoagulation and relation to occult malignancies, whereby initiation of anticoagulation unveils the presence of pre-existing cancer. The study found that anticoagulation therapy significantly increased the incidence of a positive fecal occult blood test (FOBT). However, positive FOBT was not related to increasing PT/PTT [16]. This suggests occult bleeding in anticoagulated patients was due to underlying organic lesions rather than systemic anticoagulation. With that being said, the most common causes of occult bleeding are seen in Fig. 15.8.

Fig. 15.8
figure 8

Causes of occult GI bleeding

Full size image

Management of Occult GI Bleeding on Anticoagulants

While many practitioners test FOBT to gauge a patient’s risk of GI bleeding prior to starting therapeutic anticoagulation, this is not recommended [17]. First off, FOBT has only been validated for colorectal cancer screening and not a means to evaluate GIB risk. Moreover, FOBT is an imperfect test, known for false positives in individuals consuming certain peroxidase-containing foods and false negatives in ascorbic acid-containing food items. Thus, it should not be used to predict bleeding risk prior to starting anticoagulation or used to dictate the discontinuation of anticoagulants.

In contrast, tools such as the HAS-BLED score, HEMORR2HAGES and ATRIA can be used to predict and assess a patient’s risk of GIB on anticoagulation. HAS-BLED has been favored as it more reliably predicts clinically significant GIB [20]. HAS-BLED takes into account hypertension, renal and liver dysfunction, bleeding disorder, labile INR, age greater than 65, concurrent alcohol, NSAID or aspirin use.

Certain prevention strategies can be employed however in minimizing GI bleeding from angiodysplasias, pre-existing ulcers, portal hypertensive gastropathy and colopathy and underlying malignancy. For example, adding proton pump inhibitor may mitigate the risk of GIB especially from pre-existing ulcers. Target INR for warfarin should be maintained within range and dosage of DOACs should be renally adjusted. The 2008 ACCF/ACG/AHA consensus statement recommends keeping the INR between 2.0 and 2.5 for patients (excluding high risk patients such as those with mechanical valves) on warfarin plus aspirin and clopidogrel [21]. Combination therapy with other antiplatelets and NSAIDs should be minimized and prescribed for the shortest duration indicated. As seen in Fig. 15.7 above, serum concentration of DOACs can be notoriously increased by azoles and protease inhibitors that inhibit CYP3A4 and therefore should be dose-adjusted accordingly when applicable.

Resumption of Anticoagulants after GIB

In the setting of normal renal function, DOACs are generally out of systemic circulation in approximately 12–24 h. A retrospective study by Sengupta et al. showed that restarting DOACs after GI bleeding was not associated with recurrent GI bleeding [20]. Currently, there are no formal set guidelines on resumption of DOACs. Figure 15.9 below outlines the following recommendations taken from 2016 ASGE guidelines:

Fig. 15.9
figure 9

2016 ASGE guidelines on resumption of anticoagulation after GI bleeding

Full size image