Gastrointestinal (GI) bleeding is one of the most frequent gastroenterological conditions that require medical attention. Its incidence varies with age, with older patients being more frequently affected. The severity of gastrointestinal bleeding can vary from a mild form to a potentially threatening life condition. The estimated prevalence for overall GI bleeding is approximately 3.1% (1). The origin of the bleeding defines its clinical presentation and definition, with hemorrhages originating before the Treitz ligament being classified as upper GI bleeding and those originating after this landmark as lower GI bleeding.

Upper GI bleeding in the older population is frequently associated with gastric and duodenal ulcer or esophagitis, being responsible for the episode in 80% of the cases (2). The incidence of upper GI bleeding can also vary depending on the geographical region. In Northern Europe, the annual incidence ranges from 213 to 570 per 100,000 patients (3); in a UK cohort of older patients affected with acute upper GI bleeding, the rate was 63% in those above 60 years of age (4); and in a North American cohort, this rate ranged between 35 and 45% (5). Although the incidence of upper GI bleeding in older patients is high, there is an even higher incidence of lower GI bleeding (6).

The incidence of hospitalizations for upper GI bleeding increased with age, being 197.4 per 100,000 population between 66 and 75 years of age and rising to 425.2 per 100,000 in people older than 75 years in a North American Cohort (7). In addition, the risk of readmittance after a first hospitalization remains higher than that of the general population.

As the risk of complications increases, so does the risk of mortality in a patient aged over 60 years experiencing an upper GI bleeding episode, estimated to be between 12 and 25%, compared to the 10% of the general population (2). In a study based on the Welsh population, the case fatality for an upper GI bleeding ranged from 11.2 to 21.5% in men above 65 years of age and from 9.1 to 20.7% in women, increasing with age (8).

Lower GI bleeding is a condition that most frequently appears in older patients, with an incidence that also depends on geographic and socio-economic factors and comorbidities (3). The most common etiologies are diverticulosis, ischemic colitis, colitis, hemorrhoids, and colorectal cancer (3). In Northern Europe, the incidence rate for lower GI bleeding ranges from 2.41 to 3.64 in male patients and from 1.72 to 3.10 in female patients, increasing with age (9). In a Spanish study, the incidence of new lower GI bleeding was between 100 and 150 per 100,000 patients in 2005 (6).

With the higher frequency of comorbidities, this cohort of patients also has an increased risk of hospitalization and longer in-hospital stay (10), with an estimated rate between 127.7 and 380.1 per 100,000 population, increasing with age (7). A multicentric study in a European cohort shows that this cohort of patients also suffers from an increased mortality rate, where the hospital mortality for lower GI bleeding was 2.5 and 1.17% in the following 3 months (1).

Medication history should be carefully reviewed upon admission, as many drugs may be associated with gastrointestinal bleeding.

First, anticoagulants and antiplatelets increase bleeding risk (11, 12). Their prescription is very common, as it has been reported that about half of the patients presenting for an upper gastrointestinal bleeding (UGIB) were treated using antithrombotic drugs (13). Clinicians should assess the risk–benefit ratio of antithrombotic administration in this setting. Indeed, as shown by the ASPREE trial enrolling 19,114 healthy older people (>65 years) patients without previous cardiovascular disease, the administration of low-dose aspirin increased the risk of major bleeding (hazard ratio [HR] 1.38, 95% confidence interval [95%CI] 1.18–1.62) and UGIB (HR 1.87, 95%CI 1.32–2.66) without an improvement of cardiovascular disease prevention (14). Furthermore, recent European guidelines advise that the use of aspirin for primary cardiovascular prophylaxis should be discontinued in patients who have a confirmed UGIB, although it should be continued for secondary prevention (15). Similarly, people taking warfarin should have the medication stopped along with anticoagulant reversal in situations of severe UGIB, while continued anticoagulation must be evaluated in cases of less severe UGIB (15).

Second, non-steroidal anti-inflammatory drugs (NSAIDs) are commonly prescribed in adult people; they represent a common risk factor for gastric hemorrhage. Their primary mechanism is the inhibition of cyclooxygenase 1 (COX-1); as a consequence, there is a great reduction of prostaglandin production, which leads to poorer protection of gastric mucosa. Coadministration of aspirin and other NSAIDs have been associated with greater gastric damage because of combined inhibition of both COX-1 and COX-2 mucosae-protective pathways (16).

Beyond these medications, other drugs have been associated with a potentially increased risk of GI bleeding, such as serotonin-selective reuptake inhibitors and calcium channel blockers; therefore, patients should be questioned whether they have been prescribed these drugs as well (17, 18).

Moreover, drugs per se may induce upper bleeding by causing pill esophagitis. A large number of medications have been associated with this phenomenon, with antibiotics, NSAIDS, and bisphosphonates being the most common (19). This occurrence is more frequent in older adults with a reduced esophageal transit; therefore, it should always be excluded in case of retrosternal pain, odynophagia, and drug assumption before sleeping up to 3 days before and whether the real cause of bleeding has not been identified yet (19).

Finally, the patient should be asked about recent consumption of products that may change stool appearance and make it look similar to real bleeding (i.e., iron, bismuth, liquorice, and charcoal), in order to avoid a wrong diagnosis in the initial stages (20, 21).

Upper gastrointestinal bleeding (UGIB) is defined as hemorrhage proximal to the Treitz ligament involving the esophagus, stomach, and duodenum (54). The most common symptoms of UGIB are melena, hematemesis, and coffee ground vomiting (54). Hematochezia, instead, is a rare manifestation of UGIB and is commonly a presentation of lower gastrointestinal bleeding (LGIB) (55, 56). Systemic manifestations, in major and life-threatening gastrointestinal (GI) bleeding (both UGIB and LGIB), include hemodynamic instability, hypotension, abdominal pain associated with lethargy, fatigue, syncope, and angina (56, 57).

The most common cause of UGIB is peptic ulcer disease, involving approximately 32–36% of all hospitalized patients; then esophagitis, duodenitis, and gastritis (until 24% of hospitalization); and finally variceal bleeding (approximately 11% of hospitalization, but 90% of UGIB in patients with liver cirrhosis) (4, 56, 58, 59).

The incidence of UGIB is widely different among countries ranging between 67 and 172/100.00 person with similar rates between Europe and the United States (59–61). Although hospitalizations for UGIB have declined due to H. pylori eradication, the use of proton pump inhibitors, and increased access to endoscopy (59, 62, 63), the mortality rate of UGIB is approximately 2–10% (54).

Antiplatelet and anticoagulant use, non-steroidal inflammatory drugs, corticosteroids, liver cirrhosis, the presence of multiple comorbidities, and older age are common risk factors for gastrointestinal bleeding, especially of UGIB (1, 6, 56, 62).

Several guidelines suggest general recommendations for initial management of UGIB (a flowchart of UGIB management is proposed in Figure 1). First, patients with UGIB should be guaranteed an IV access by cannula (≤18 G) in each antecubital fossa and an early fluid resuscitation (37) should be started, reducing the risk of mortality and myocardial infarction (54), achieving 90–100 mmHg systolic blood pressure as target (64, 65). In particular, a first approach with 500 mL of crystalloids infused in less than 15 min are suggested as first choice in hemodynamically unstable patients (66), although studies showed no difference between colloids and crystalloids in fluid resuscitation during UGIB (54).

During resuscitation, transfusion of packed red cells should be performed with a restrictive approach using a threshold of 7 g/dL (<8 g/dL in patients with cardiovascular disease) for hemoglobin (Hb) (37, 54) and transfusion of platelets should be performed using a threshold of 50 × 10⁹/L (67, 68). After resuscitation, treatment of UGIB is divided into three phases: pre-endoscopic, endoscopic, and post-endoscopic phases.

Lower gastrointestinal bleeding (LGIB) represents the 3% of emergency surgical referrals (78), and its incidence is estimated to be 33–87 for 100.000 patients (79). The mortality is 3.4% rising 18–20% in patients with LGIB during hospitalization (79). The most common cause of LGIB is diverticular bleeding, followed by benign anorectal conditions such as hemorrhoids, fissures, and rectal ulcers (79). Other common causes are telangiectasia in multiple sites of GI tract, colitis, and colorectal cancer. Of note, 23% of hospitalized patients with LGIB in the UK are discharged without a diagnosis (80). Patients with LGIB should be clinically evaluated to establish the hemodynamic stability. Clinical history (bleeding history and comorbidities), clinical evaluation (including digital-rectal exploration), laboratory test, and concomitant therapy are needed to establish the hemodynamic status of patients (81). In particular, shock index (heart rate [HR] and systolic blood pressure [SBP] ratio) is recommended by current guidelines (79). A shock index >1 defined the patient as hemodynamically unstable (79).

Similar to UGIB, a resuscitation strategy (previous described) should be performed in unstable patients (79). In these patients, a computed tomography (CT) scan with angiography should be performed to evaluate the focus of bleeding; then, patients should undergo to interventional radiology (preferably <60 min from hospital admission) or endoscopy (79). Although endoscopy treatment represents the first line of treatment from international guidelines (79, 81, 82), only 2.1% of cases of LGIB undergo endoscopic treatment and the most common intervention is red blood cell transfusions (79). If a treatment failure occurred during endoscopy or radiological intervention, surgery should be evaluated in selected cases. If no focus of bleeding is identified during CT scan with angiography, patients should be considered stable with major bleeding.

If shock index is <1, LGIB should be considered stable and Oakland score should be performed to establish whether major or minor bleeding occurred and whether hospitalization is required (Oakland score ≤ 8 suggests a possible outpatient management) (79).

Fluid resuscitation is needed, and a restricted red blood cell transfusion regimen should be preferred with a threshold of Hb <70 g/L with a target of 70–90 g/L after transfusion, except for patients with previous history of cardiovascular disease with a threshold of Hb <80 g/L with a target of 10.0 g/L (79), and platelets transfusion should be performed using a threshold of 50 × 10⁹/L (82). Figure 2 summarizes the management of LGIB.

In major bleeding, if patient is hemodynamically stable, colonoscopy should be performed after adequate colon cleansing (a nasogastric tube may help colon preparation in patients with a low risk of aspiration and ongoing bleeding) (82). In total, 4–6 liters of a polyethylene glycol (PEG)-based solution or the equivalent should be administered over 3–4 h until the rectal effluent is clear of blood and stool (82).

In patients with high-risk clinical features or ongoing bleeding, endoscopy should be performed within 24 h of patient’s admission to the emergency department, after an adequate colon preparation; otherwise, a colonoscopy should be performed next available after a colon purge (82). Further investigations are needed if no bleeding focus was found, such as CT scan with angiography repetition or use of video capsules (79).

In older people, the severity and prognosis of GI bleeding are influenced by medical comorbidities (1) and therapies as well as the use of antiplatelet and anticoagulants medication (83). Indeed, 70% of UGIB occurred in patients >60 years old and its incidence and mortality risk rise according to age (83). Similarly, patients with LGIB are more common in older patients, with a mean onset age between 63 and 77 years, with higher mortality risk (83). In addition, GI bleeding incidence seems to be reduced only in patients <70 years old (61). For these reasons, GI hemorrhage management is a backbone in the older care and there are some peculiarities of old age that should be addressed.

First, endoscopy, the first line of diagnosis and treatment for GI hemorrhage, has similar mortality risk in older patients compared to general population and old age is not a contraindication to endoscopy (83). Of note, older patients had an increased risk to developing adverse events and oxygen desaturation, especially if benzodiazepines (BDZ) are administered during endoscopy for sedation (83, 84); for this reason, a lower dose of BDZ with careful titration is suggested (83).

In addition, older patients are more likely to be treated with antiplatelets, anticoagulants, especially with complex antithrombotic therapy (CAT) resulting in an increased risk of hospitalization and transfusion as shown in a large cohort study of 78,133 old veterans aged >60 years treated with antiplatelets and/or anticoagulants, with the highest risk of hospitalization and transfusion in patients treated with dual antiplatelet agents and anticoagulant (85).

In patients treated with antiplatelets, a GI bleeding incidence rate of 0.7–1.3% for aspirin and 1.2–2% for aspirin and clopidogrel combination during a follow-up of 1–2 years was observed in Western countries (86–88). In older patients, the GI bleeding incidence rate rise to 2.7% as shown by an observational study on 1852 patients undergone the implantation of drug-eluted stent (DES) with a mean age of 70.9 years (89).

Antiplatelet management had a fundamental role in older patients. If antiplatelet is administered in primary prophylaxis should be permanently discontinued, while antiplatelet in secondary prevention should not be stopped, but if suspension is needed, it should be restarted when hemostasis is guaranteed (79, 82).

In patients treated with dual antiplatelet therapy with aspirin and a P₂Y₁₂ inhibitor, P₂Y₁₂ should be stopped only in unstable hemorrhage and restarted within 5 days, especially if recent coronary stenting is performed (54, 79, 81).

In older patients, comorbidities such as atrial fibrillation (AF) that require indefinite anticoagulation are common. The most common oral anticoagulant prescribed in older patients is vitamin K antagonists (VKAs) and direct oral anticoagulants (DOACs).

In older patients (≥75 years), the incidence rate of GIB was 2.19% per year for dabigatran 110 mg and 2.80% per year for dabigatran 150 mg (90). Instead, an incidence GIB rate of 1.51 and 0.83% per year was observed for edoxaban 60 mg and 30 mg, respectively (90). The incidence of GIB rate of 2.0% for rivaroxaban and 0.76% for apixaban, but no data according to dose were found (90). The incidence of GIB for warfarin ranged between 0.86 and 1.59% per year in phase III clinical trials (90). A large cohort meta-analysis performed on 129.357 patients by Miller et al. summarized the risk of GIB in patients treated with DOACs and VKAs (pooled rate: 1.5% versus 1.3%, respectively; odds ratio [OR]: 0.98; 95%CI: 0.80–1.21) (91).

Recently, a large network meta-analysis on 605.771 AF patients showed a reduced risk of GIB in patients treated with apixaban compared to ones treated with dabigatran or rivaroxaban (92).

VKAs, as warfarin, should be discontinued during hemorrhage and should be restarted at 7 days after GI bleeding if low thrombotic risk (54, 79, 81). In patients with high thrombotic risk as well as patients with mechanical prosthetic heart valve, AF with prosthetic heart valve or mitral stenosis, <3 months after venous thromboembolism, low molecular weight heparin treatment should be started at 48 h after hemorrhage (54, 79). In patients on warfarin, although the optimal international normalized ratio (INR) to perform endoscopy is <1.3, endoscopy may be also considered when INR is <2.5 without a significantly increased risk of rebleeding (54).

On the other hand, DOAC should be stopped at presentation of GI bleeding and restarted within 7 days after GI bleeding (54, 79, 81).

In life-threatening GI bleeding and hemodynamically unstable patients, the interruption of oral anticoagulants is not enough, and a reversal agent is needed (Figure 3) (12, 54, 79, 81).

In VKAs, Vitamin K is a specific reversal agent in a dose-dependent manner (1 to 10 mg). Slow intravenous administration (in 25 to 50 mL normal saline over 15 to 30 min) causes a rapid reduction in the INR (4–6 h) (93). However, the administration of vitamin K does not result in immediate correction of coagulopathy, and in life threatening bleeding, vitamin K administration must be accompanied by the administration of 4-factor prothrombin complex concentrate (4F-PCCs), or, if not available, plasma (93). 4F-PCCs should be administered according to INR range and body weight (INR 2–4 at 25 U/kg, INR 4–6 at 35 U/kg, and INR >6 at 50 U/kg; max dose 5,000 U if 100 kg body weight) (93).

In DOACs, in dabigatran users, a reversal agent, idarucizumab is suitable (5gr + 2.5gr, IV) (93). However, in patients taking apixaban, rivaroxaban, and edoxaban, andexanet alfa may be useful. If these drugs are not available, 4F-PCC or activated PCC (aPCC) may be an alternative (50 U/kg) (12, 93).

Furthermore, a large cohort study of 3,166 patients treated with antiplatelet and without routine PPI use, due to previous myocardial infarction or cerebrovascular event, showed that the long-term risk of bleeding is higher in older patients than in younger patients with a substantial risk of disabling or fatal UGIB, suggesting that a co-prescription of PPI should be encouraged (94).

On the other hand, the use of PPI is associated with several adverse effects, as well as increased risk of fractures, osteoporosis, higher risk of Clostridium difficile (CD) infection and community-acquired pneumonia (CAP), especially in older patients (95).

In particular, a large meta-analysis including 2,181,546 individuals taking or not taking PPI showed that patients not taking PPI, those taking PPI, had an increased risk of developing any-site fractures, hip fractures, spine fracture, and osteoporosis (96), this evidence is confirmed independently from dose and duration of therapy, as suggested by a large meta-analysis on 2,103,800 patients showing a high risk of hip fracture in patients with long- and short-term therapy and in low, medium, and high dosage of PPI (97).

In addition, PPI is also associated with an increased risk of developing CD infection as shown in a large meta-analysis of 56 studies (40 case control and 16 cohort) involving 356,683 patients (98). This risk is estimated approximately 64% compared to ones not taking PPI (99) and probably is related to PPI-gut dysbiosis (100, 101) that increases also all-cause mortality (101).

Finally, several studies showed an increased risk of developing CAP in patients taking PPI compared to ones who are not taking these drugs, particularly within 30 days (102–104). A pathogenic mechanism has been proposed to explain the association between PPI use and the incidence of CAP: PPIs may increase the gastric pH altering also normal oropharyngeal flora, which could increase susceptibility to respiratory infections by permitting survival of pathogens that lead to CAP (104).

Osteoporosis, bone fractures, CD infection, and CAP may be deadly for older patients, and these complications should be avoided. For this reason, PPI in older patients should be used only according to clinical indications, with long-term treatments only for selected cases (36).

In conclusion, UGIB and LGIB represent a severe common complication, especially in older patients with comorbidities and on treatment with antiplatelet and/or anticoagulant drugs. Several drugs are available to reduce bleeding complications, especially for UGIB. Current evidence and guidelines suggest a clinical approach based on hemodynamic status with endoscopy as the first line for diagnosis and treatment.

DM: Conceptualization, Writing – original draft. GG: Writing – original draft. FS: Writing – original draft. AP: Writing – review & editing. PP: Supervision, Writing – review & editing. DP: Conceptualization, Supervision, Visualization, Writing – review & editing.