bs_bs_banner
Journal of Digestive Diseases 2014; 15; 283–292
doi: 10.1111/1751-2980.12140
Original article
Opposing effects of aspirin and anticoagulants on morbidity and mortality in patients with upper gastrointestinal bleeding Hussein ABU DAYA,*,† Mohamad ELOUBEIDI,† Hani TAMIM,‡,§ Houssam HALAWI,‡ Ahmad H. MALLI,‡ Don C. ROCKEY¶ & Kassem BARADA†
¶
*Department of Internal Medicine, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, Department of Internal Medicine, Medical University of South Carolina, Charleston, South Carolina, USA; and †Division of Gastroenterology and Hepatology, ‡Department of Internal Medicine, and §Biostatistics Support Unit, Clinical Research Institute, American University of Beirut Medical Center, Beirut, Lebanon
OBJECTIVE: We aimed to determine the effect of antithrombotics on in-hospital mortality and morbidity in patients with peptic ulcer disease-related upper gastrointestinal bleeding (PUD-related UGIB). METHODS: The study cohort was retrospectively selected from a tertiary center database of patients with PUD-related UGIB, defined as bleeding due to gastric or duodenal ulcers, or erosive duodenitis, gastritis or esophagitis. Outcomes were compared among patient groups based on their antithrombotic medications before admission. Patients on no antithrombotics served as controls. The composite adverse outcomes, in-hospital mortality, rebleeding and/or need for surgery were measured. Severe bleeding and in-hospital complications were also recorded. RESULTS: Of 398 patients with PUD-related UGIB, 44.5% were on aspirin or anticoagulants only. The
composite adverse outcome was most common in patients taking anticoagulants only (40.5%), intermediate in controls (23.1%) and least in those taking aspirin only (12.1%). On multivariate analysis, patients taking aspirin alone had a significantly lower risk of adverse outcome events (odds ratio [OR] 0.4, 95% CI 0.2–0.8) and a shorter length of hospital stay (regression coefficient = −3.4, 95% CI [−6.6, −0.6]). In contrast, taking anticoagulants was associated with a greater risk of adverse outcome events (OR 2.3, 95% CI 1.0–5.3), severe bleeding (OR 2.6, 95% CI 1.2–5.8) and in-hospital complications (OR 2.9, 95% CI 1.3– 6.6). CONCLUSIONS: Patients with PUB-related UGIB while taking aspirin had fewer adverse outcomes compared with those taking anticoagulants. Aspirin may have beneficial effects in this population.
KEY WORDS: anticoagulant, gastrointestinal hemorrhage, morbidity, peptic ulcer hemorrhage, platelet aggregation inhibitor.
Correspondence to: Kassem BARADA, Division of Gastroenterology and Hepatology, American University of Beirut Medical Center, Riad El Solh, Beirut 1107 2020, Lebanon. Email: [email protected] Conflicts of interest: None. © 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
INTRODUCTION Upper gastrointestinal bleeding (UGIB) is associated with high morbidity and mortality.1 In the USA the annual estimated hospitalization nationwide due to peptic ulcer disease (PUD) ranged from 156 108 to 222 601, with a decreasing trend from 1993 to 2006.2 Up to two-thirds of UGIB admissions are due to PUD.3
283
284
H Abu Daya et al.
Despite the development of treatment modalities, the 30-day mortality in patients suffering from UGIB was reported to be 8.9% in the UK,4 6.2% in Hong Kong SAR, China and 8.5% in the USA.5,6 Antiplatelet agents and anticoagulants are increasingly applied in the clinic setting for the treatment and prophylaxis of cardiovascular diseases, especially in elderly populations.7 Monotherapy with antiplatelet agents or anticoagulants is associated with an increasing risk of UGIB8 and aspirin has been verified to play an important role in the pathogenesis of PUD. Therefore, it is common practice to discontinue these medications in patients with UGIB for days, even weeks, during and after bleeding episode.9 Although the continuation of low-dose aspirin will increase the risk of rebleeding, it has been reported to be able to reduce the all-cause mortality rate (1.3% vs 12.9% compared with placebo) in a small sample of patients.10 Current evidence on the efficacy of aspirin or anticoagulants on the clinical outcomes of patients with UGIB is controversial, as some authors10–13 have reported that aspirin decreases the mortality of patients while others1,14,15 have suggested these drugs have no effects. Similarly, conflicting results on the effects of anticoagulants such as warfarin have been reported.11,12,16–18 These drugs have been reported to increase the patients’ mortality significantly13 or having no influence on the disease course of UGIB, except that they are associated with the patients’ prolonged hospital stay.15 In this study we aimed to determine the impact of aspirin and anticoagulants on the clinical outcomes of patients who were admitted to hospital and hospitalized due to PUD-related UGIB, and to investigate the etiology of death in patients who had a fatal outcome. PATIENTS AND METHODS Patients Patients who were admitted to hospital and hospitalized in the Division of Gastroenterology and Hepatology, American University of Beirut Medical Center (Beirut, Lebanon) from 1993 to 2010 with evidence of gross UGIB were identified using the International Statistical Classification of Diseases and Related Health Problems (ICD)-9/ICD-10 coding system. UGIB was defined as having the symptoms and signs of reported or witnessed hematemesis, melena or hematochezia with an identified source of
Journal of Digestive Diseases 2014; 15; 283–292 bleeding in the upper gastrointestinal tract by esophagogastroduodenoscopy (EGD). PUD was considered to be present in those with gastric or duodenal ulcers or esophageal, gastric or duodenal erosions. All patients diagnosed with PUD-related UGIB during the study period were recruited in the study. Exclusion criteria were: (i) non-PUD-related UGIB; (ii) small intestinal or lower gastrointestinal bleeding, defined as hematochezia or melena with the bleeding source identified by colonoscopy or small intestinal followthrough, push enteroscopy, ileoscopy or computed tomography examination of the abdomen and pelvis, respectively, but in the absence of UGIB lesion on EGD; and (iii) occult UGIB. The study was approved by the Institutional Review Board. Antithrombotic medications A history of antithrombotic medications used at the time of diagnosis was obtained by reviewing the medical records of the patients as well as the notes from the nurses and physicians that contained sections devoted to the patients’ intake of medications at home. The patients were then divided into five groups based on their drug intake before admission: (i) those who had not taken any anticoagulants or antiplatelet agents (group A); (ii) those taking anticoagulants (warfarin, heparin or low-molecularweight heparin) only (group B); (iii) those taking aspirin only (group C); (iv) those taking any antiplatelet agent except aspirin (clopidogrel, ticlopidine or dipyridamole) with or without aspirin (group D); and (v) those taking antiplatelet agents combined with anticoagulants (group E). Study variables Patients’ characteristics including their individual variables such as age and gender, past medical history, clinical manifestations, vital signs and blood test results when they had been taken to the Emergency Room, their management in the Emergency Room and in hospital, diagnostic EGD findings, types of therapeutic endoscopic procedures, angiography and embolization, if any, and surgical treatment, if any, as well as their in-hospital mortality. The etiologies of in-hospital mortality were documented and divided into the following types: (i) cardiovascular/ thromboembolic events such as myocardial infarction, pulmonary embolus and stroke; (ii) a defined focus of infection associated with septic shock; (iii) death due to uncontrolled gastrointestinal bleeding or
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 283–292
Antithrombotics on UGIB outcomes
complications related to the inventions used to control bleeding; (iv) systemic cancer; and (v) multiorgan failure. Definition of clinical outcomes Primary or composite adverse outcomes were defined as the occurrence of at least one of the following factors: in-hospital mortality, need for surgery (requiring any kinds of surgical procedures to control bleeding) and rebleeding. Secondary outcomes were defined as severe bleeding, in-hospital complications and the length of hospital stay. The need for and the amounts of blood transfusion given were recorded. Rebleeding was defined as the recurrence of hematemesis, coffee ground emesis or melena 24 h after the initial endoscopic evaluation or hemostatic therapy and initial stabilization, accompanied by either a change in vital signs or a decrease in hemoglobin level by at least 20 g/L. Rebleeding events during the same hospitalization or requiring another hospitalization were combined upon the retrieval of data in patients with ICD-9/10 codes of UGIB. In the case of multiple admissions, only the first admission was included and subsequent re-admissions were considered as rebleeding episodes if they took place within one month of the initial admission. Severe bleeding was defined based on the following criteria: (i) hypotension: systolic blood pressure < 90 mmHg at admission; (ii) tachycardia: >119 beats/min at admission; or (iii) transfusion of more than 750 mL packed red blood cells. In-hospital complications included cardiovascular or thromboembolic events (myocardial infarction, angina, deep vein thrombosis, pulmonary embolism, stroke and transient ischemic attack, etc.), infectious complications including pneumonia, urinary tract infection, skin infections, sepsis and other complications such as acute respiratory distress syndrome (ARDS), renal failure, the need for mechanical ventilation and disseminated intravascular coagulation. Statistical analysis Statistical analyses were performed using SAS 9.1 (SAS Institute, Cary, NC, USA). Categorical variables were expressed as numbers and percentages, and continuous data were expressed as mean ± standard deviation. Bivariate analyses were performed using the χ2 test or Student’s t-test, where appropriate. To control for the effect of potentially confounding variables multivariate analyses were carried out. For
285
Gastrointestinal bleeding (n = 1077)
LGIB (n = 220)
NV-UGIB (n = 717)
Non-PUD-related UGIB‡ (n = 319)
Admissions excluded† (n = 140)
PUD-related UGIB (n = 398)
Figure 1. Flowchart of patients’ enrollment and grouping in patients with peptic ulcer disease (PUD)-related upper gastrointestinal bleeding (UGIB). †Esophageal variceal bleeding (n = 85), small intestinal bleeding (n = 9) and occult gastrointestinal bleeding (n = 46). ‡Causes of nonPUD-related include Mallory–Weiss tear, arteriovenous malformation, Dieulafoy’s lesion, malignancy and hiatal hernia. LGIB, lower gastrointestinal bleeding; NV, non-variceal.
categorical outcomes (such as the composite adverse outcomes), a multivariate logistic regression was performed and the odds ratio (OR) and 95% confidence interval (CI) were calculated. On the other hand, for continuous outcomes (such as the length of hospital stay), multivariate linear regression was carried out where the β coefficient and 95% CI were recorded. Variables included in the regression model were those of either statistical or clinical significance. P ≤ 0.05 was considered statistically significant. RESULTS Characteristics of the patients Altogether 1077 patients with acute gastrointestinal bleeding were diagnosed and treated in our hospital from 1993 to 2010. Of these patients, 679 were excluded from the study for lower gastrointestinal bleeding (n = 220), esophageal variceal bleeding (n = 85), small intestinal bleeding (n = 9), occult gastrointestinal bleeding (n = 46) and non-PUD-related UGIB (n = 319). Finally, 398 patients aged 65.1 ± 15.2 years were enrolled in the study (Fig. 1), with a significant male predominance (male vs female: 73.9% vs 26.1%; Table 1). Patients with PUD-related UGIB in the group A were significantly younger and had less comorbidity than those taking antithrombotics (P < 0.01). The proportion of patients using proton pump inhibitor (PPI) before presentation was similar among the different groups (8% in group A vs 7.1% in group C [P = 0.80] and 16.2% in group B [P = 0.10], respectively). The frequency of patients having alcohol intake was significantly lower in group B than in group
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
11 (29.7) 16 (43.2) 25 (67.6) 17 (45.9)***
51 (31.9) 72 (45.0) 115 (71.9) 83 (51.9)
5 (13.5) 17 (45.9) 11 (29.7) 2 (5.4) 4 (10.8)
35 (21.9) 81 (50.6) 29 (18.1) 4 (2.5) 26 (16.3) 0.24 ± 0.06 2.9 ± 1.8**
12 (32.4)** 26 (70.3)** 18 (48.6)** 10 (27.0)** 7 (18.9) 7 (18.9)
21 (13.1) 50 (31.3) 19 (11.9) 2 (1.3) 29 (18.1) 39 (24.4)
0.27 ± 0.08 1.2 ± 0.7
71.1 ± 11.0** 21 (56.8)
60.5 ± 18.3 112 (70.0)
Group B (n = 37)
55 (39.3) 60 (42.9) 105 (75.0) 88 (62.9)***
0.27 ± 0.06 1.0 ± 0.2*
21 (15.0) 77 (55.0) 27 (19.3) 5 (3.6) 28 (20.0)
45 (32.1)** 76 (54.3)** 50 (35.7)** 4 (2.9) 13 (9.3)* 23 (16.4)
66.9 ± 12.0** 110 (78.6)
Group C (n = 140)
13 (38.2) 14 (41.2) 25 (73.5) 20 (58.8)
0.26 ± 0.04 1.2 ± 0.6
7 (20.6) 8 (23.5)** 12 (35.3)* 4 (11.8)** 11 (32.4)*
11 (32.4)** 25 (73.5)** 26 (76.5)** 0 (0) 5 (14.7) 3 (8.8)*
70.0 ± 13.0** 28 (82.4)
Group D (n = 34)
8 (29.6) 8 (29.6) 15 (55.6) 17 (63.0)
0.26 ± 0.07 2.2 ± 2.0**
4 (14.8) 16 (59.3) 6 (22.2) 2 (7.4) 5 (18.5)
9 (33.3)** 16 (59.3)** 21 (77.8)** 6 (22.2)** 2 (7.4) 0 (0)**
70.0 ± 9.0** 23 (85.2)
Group E (n = 27)
138 (34.7) 170 (42.7) 285 (71.6) 225 (56.5)
0.27 ± 0.07 1.5 ± 1.1
72 (18.1) 199 (50.0) 85 (21.4) 17 (4.3) 74 (18.6)
98 (24.6) 193 (48.5) 134 (33.7) 22 (5.5) 56 (14.1) 72 (18.1)
65.1 ± 15.2 294 (73.9)
Total (n = 398)
H Abu Daya et al.
*P < 0.05, **P < 0.01, ***P = 0.054 compared with group A. †Each symptom and endoscopic diagnosis (including patients with more than one symptom) are included. ‡Patients with gastric ulcer and/or duodenal ulcer. Univariate analysis is performed with group A (those who are not taking any antithrombotics) as the control group. CAD, coronary artery disease; Group A, patients who have taken no anticoagulants or antiplatelet agents; group B, patients taking anticoagulants (warfarin, heparin or low-molecular-weight heparin) only; group C, patients taking aspirin only (group C); group D, patients taking any antiplatelet agents except aspirin (clopidogrel, ticlopidine or dipyridamole) with or without aspirin; group E, patients taking antiplatelet agents combined with anticoagulants; INR, international normalized ratio; NSAID, non-steroidal anti-inflammatory drugs; PMH, past medical history; SD, standard deviation.
Age, years (mean ± SD) Male gender, n (%) PMH, n (%) Diabetes Hypertension CAD Atrial fibrillation Cancer Non-aspirin NSAIDs use Symptoms at presentation, n (%)† Hematemesis Melena Hematemesis + melena Hematochezia Syncope Initial laboratory results Hematocrit (mean ± SD) INR (mean ± SD) Endoscopic diagnosis, n (%)† Gastric ulcer Duodenal ulcer Peptic ulcer‡ Erosion
Group A (n = 160)
Table 1. Individual and clinical characteristics of patients with peptic ulcer disease-related upper gastrointestinal bleeding, comparing those taking different antiplatelet agents and anticoagulants
286
Journal of Digestive Diseases 2014; 15; 283–292
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
287
*P < 0.05; **P < 0.01, and ***P = 0.056 compared with the group A. †Composite adverse outcomes are defined as the occurrence of in-hospital mortality, need for surgery or rebleeding. Group A, patients have not taken any anticoagulants or antiplatelet agents; group B, those taking anticoagulants (warfarin, heparin or low-molecular-weight heparin) only; group C, those taking aspirin only (group C); group D, those taking any antiplatelet agents except aspirin (clopidogrel, ticlopidine or dipyridamole) with or without aspirin; and group E, those taking antiplatelet agents combined with anticoagulants.
16 (4.0) 24 (6.0) 66 (16.6) 79 (19.8) 162 (40.7) 136 (34.2) 96 (24.1) 301 (75.6) 1250 ± 2500 7±9 0 (0) 0 (0) 4 (14.8) 4 (14.8) 16 (59.3)* 15 (55.6)** 7 (25.9) 23 (85.2)* 1250 ± 1250 10 ± 11
Total (N = 398) Group E (n = 27) Group D (n = 34)
1 (2.9) 0 (0) 6 (17.6) 6 (17.6) 9 (26.5) 10 (29.4) 11 (32.4) 30 (88.2)** 750 ± 500 5±4 1 (0.7)** 4 (2.9)* 14 (10.0)* 17 (12.1)* 53 (37.9) 39 (27.9) 35 (25.0) 109 (77.9)* 875 ± 750* 5 ± 4* 4 (10.8) 5 (13.5) 12 (32.4)*** 15 (40.5)* 24 (64.9)** 24 (64.9)** 8 (21.6) 35 (94.6)** 2750 ± 6500* 12 ± 15***
Mortality (0.7% vs 6.3%, P < 0.01) and rebleeding rate (10.0% vs 18.8%, P < 0.05) were less frequent in the patients treated with aspirin only (group C) than those taking neither antiplatelet agents nor anticoagulants (group A), while those taking anticoagulants only (group B) had both the highest mortality and the highest rebleeding rate among all the five groups (10.8% and 32.4%, respectively) (Table 2). In addition, the length of hospital stay was significantly shorter in patients in group C than in groups A, B and E, but was similar to that in group D. A high proportion of patients in group C had a blood transfusion; however, they were given fewer volumes of blood transfusion per patient than group A (875 ± 750 mL vs 1250 ± 1750 mL, P < 0.05). In contrast, patients in group B were most likely to require a blood transfusion and received the highest volumes of blood transfusions, and they had the longest length of hospital stay among all the groups, although this was of borderline significance (P = 0.056). Similar overall associations were observed in patients in group D compared with the group A, as they had lower mortality (2.9% vs 6.3%), less need for surgery (0% vs 9.4%) and composite adverse outcome rates (17.6% vs 23.1%) together with a shorter duration of hospital
10 (6.3) 15 (9.4) 30 (18.8) 37 (23.1) 60 (37.5) 48 (30.0) 35 (21.9) 104 (65.0) 1250 ± 1750 8 ± 11
Clinical outcomes of the patients
Mortality n (%) Surgery n (%) Rebleeding n (%) Composite adverse outcomes† n (%) Severe bleeding n (%) In-hospital complications n (%) Endoscopic therapy n (%) Blood transfusion n (%) Blood transfusion, mL (mean ± SD) Length of hospital stay, days (mean ± SD)
In total, 285 (71.6%) patients were found to have peptic ulcers (gastric and/or duodenal ulcers) and 225 (56.5%) had erosions. More than 20% of the patients had multiple lesions. There was no difference in the frequency of duodenal or gastric ulcers among the groups.
Group C (n = 140)
Patients in the cohort had evidence of substantial blood loss, with a mean hematocrit level of 0.27 ± 0.07 (Table 1). Melena was the most common presentation, occurring in half the patients, followed by hematemesis combined with melena (21.4%), syncope (18.6%), hematemesis only (18.1%) and hematochezia (4.3%), respectively. The mean international normalized ratio (INR) was higher in patients who were using anticoagulants only (2.9 ± 1.8, P < 0.01) but lower in those using aspirin only (1.0 ± 0.2, P < 0.05), respectively, than those in group A who were treated with neither anticoagulants nor antiplatelet agents (1.2 ± 0.7).
Group B (n = 37)
Clinical presentations and the etiology of UGIB
Group A (n = 160)
A (18.6% vs 3.3%, P = 0.04), but it was otherwise similar in the different groups.
Antithrombotics on UGIB outcomes Table 2. Frequencies of clinical outcomes in patients with peptic ulcer disease-related upper gastrointestinal bleeding and comparison between patients taking any anticoagulants (AC) or antiplatelet (AP) and controls
Journal of Digestive Diseases 2014; 15; 283–292
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
0.0001 0.7000 0.0001 0.0001
56 (34.6) 53 (32.7) 36 (22.2) 17 (10.5) NA 53 (46.9) 24 (28.6) 65 ± 15 115 (43.9) 90 (34.3) 46 (17.6) 11 (4.2) 84 (32.1) 101 (47.4) 13 (10.4) 62 ± 15 NA 0.01 0.03 0.94
46 (33.8) 44 (32.4) 30 (22.0) 16 (11.8) 78 (57.4) 39 (44.8) 24 (33.3) 71 ± 13 128 (40.1) 111 (34.8) 58 (18.2) 22 (6.9) NA 123 (50.0) 22 (15.2) 65 ± 15
0.71
0.004 0.010 0.330 0.940
33 (41.8) 23 (29.1) 18 (22.8) 5 (6.3) NA† 17 (31.5) 15 (28.8) 65 ± 17 8 (50.0) 6 (37.4) 1 (6.3) 1 (6.3) 12 (75.0) 1 (9.1) 4 (28.6) 65 ± 17
153 (40.1) 128 (33.5) 75 (19.6) 26 (6.8) 150 (39.3) 139 (48.1) 33 (18.0) 65 ± 15
0.590
†Composite adverse outcomes (primary outcomes) include in-hospital mortality, rebleeding or need for surgery. ‡Comorbidity includes coronary artery disease, congestive heart failure, cerebral vascular accident, systemic cancer, chronic renal failure and diabetes mellitus. §Percentages calculated as: (patients with the variables [yes or no] in group C or B)/(patients with the variables [yes or no] in group C or B + those with the variables [yes or no] in group A). NA, not applicable.
NA 0.900 0.002 0.990
26 (39.4) 18 (27.3) 17 (25.7) 5 (7.6) NA 14 (31.8) 12 (28.6) 67 ± 15 105 (44.5) 81 (34.3) 40 (16.9) 10 (4.3) NA 87 (46.5) 13 (11.5) 65 ± 15
0.02 0.01
Yes P value No Yes P value No Yes P value No
NA 0.03 0.07 0.30 135 (40.7) 116 (34.9) 59 (17.8) 22 (6.6) NA 126 (49.2) 25 (16.1) 65 ± 15
0.40
P value No No
P value
Yes
Rebleeding Severe bleeding
Number of comorbidities, n (%)‡ 0 1 2 ≥3 Severe bleeding, n (%) Aspirin only, n (%)§ Anticoagulants only, n (%)§ Age, years (mean ± SD)
To further explore the protective effect of aspirin against in-hospital mortality, we determined the cause of death in the 16 patients, in which one was using aspirin only, one used a combination of aspirin and clopidogrel, three used heparin, one used coumadin and ten took no antithrombotics. Among these
Yes
On multivariate analysis, elder age (> 60 years) was found to be an independent predictor of in-hospital complications (OR 3.2, 95% CI 1.6–6.1, P = 0.001; Table 4). In addition, intake of aspirin alone at presentation was independently associated with fewer adverse outcomes (OR 0.4, 95% CI 0.2–0.8, P = 0.01) and shorter length hospital stay (r = −3.4, 95% CI [−6.6, −0.6], P = 0.02). On the other hand, being on anticoagulants was associated with a greater risk of adverse outcomes, in-hospital complications and severe bleeding (P = 0.04, 0.007 and 0.02, respectively). Finally, endoscopic therapy was an independent predictor of rebleeding (P = 0.01) and adverse outcomes (P = 0.01) in patients on anticoagulants. PPI use was not associated with the composite adverse outcomes (OR 0.7, 95% CI 0.2–2.4, P = 0.50), in-hospital complications (OR 0.9, 95% CI 0.3–2.3, P = 0.80), severe bleeding (OR 0.5, 95% CI 0.2–1.3, P = 0.20) or rebleeding (OR 0.5, 95% CI 0.1–2.3, P = 0.40).
In-hospital complications
Patients who had in-hospital complications were on average a decade elder than those who did not (P = 0.0001) (Table 3). Similarly, the frequencies of patients with in-hospital complications and severe bleeding increased as the number of their comorbidities increased (P = 0.01 and 0.02, respectively). Moreover, the frequency of severe bleeding was higher in patients who suffered in-hospital death and complications than in those who did not (P = 0.004 and 0.0001, respectively). Finally, patients who died or had composite adverse outcomes were less likely to be on aspirin only at presentation than those who did not (9.1% vs 48.1%, P = 0.010; and 31.5% vs 50.0%, P = 0.01). In contrast, those had composite adverse outcomes (P = 0.03), in-hospital complications (P = 0.0001) and severe bleeding (P = 0.002) were more likely to be on anticoagulants only than those who did not; however, those on anticoagulants were elder and had more comorbidities than group A (1.30 ± 1.00 vs 0.56 ± 0.81, P < 0.0001).
Composite adverse outcome†
stay (5 ± 4 days vs 8 ± 11 days). Interestingly, there were no deaths in patients taking both antiplatelet agents and anticoagulants (group E) although they were similar in severity of bleeding and rates of in-hospital complications to group B.
Journal of Digestive Diseases 2014; 15; 283–292
Mortality
H Abu Daya et al.
Table 3. Predictors of in-hospital mortality, composite adverse outcomes, rebleeding, severe bleeding and in-hospital complications using univariate analysis in patients with peptic ulcer disease-related upper gastrointestinal bleeding
288
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Anticoagulatn use
*P < 0.05, **P ≤ 0.01. †Multivariable analysis for aspirin use and AC use is performed with group A (those who were not taking any antithrombotics) as the control group. ‡Comorbidities include coronary artery disease, congestive heart failure, cerebral vascular accident, systemic cancer, chronic renal failure and diabetes mellitus. §Age ≤ 60 years vs >60 years.
0.01 0.80 0.07 0.06 0.14 0.9 0.5 0.01 0.2–0.8 0.6–1.4 0.9–4.6 0.9–3.9 0.8–4.6 0.7–1.6 0.6–3.2 1.2–5.5 0.4–1.2 0.9–1.7 1.6–6.1 0.6–2.1 1.3–6.6 0.9–1.8 1.0–4.5 0.5–2.2 Aspirin use
Aspirin only† Comorbidities‡ Age§ Endoscopic therapy Anticoagulants only† Comorbidities‡ Age § Endoscopic therapy
0.4** 0.9 1.3 1.7 2.3* 0.9 1.2 2.6**
0.2–0.8 0.6–1.4 0.6–2.5 0.8–3.2 1.0–5.3 0.6–1.4 0.6–2.6 1.3–5.4
0.01 0.69 0.47 0.14 0.04 0.70 0.60 0.01
0.7 1.2 3.2** 1.2 2.9** 1.3 2.1* 1.0
0.15 0.20 0.001 0.60 0.007 0.20 0.04 0.90
0.9 1.2 1.1 1.5 2.6* 1.3 1.0 1.3
0.6–1.5 0.9–1.6 0.7–1.9 0.9–2.7 1.2–5.8 0.9–1.9 0.5–1.9 0.6–2.5
0.70 0.20 0.70 0.10 0.02 0.10 0.90 0.50
0.4** 1.0 2.1 1.9 1.9 1.0 1.4 2.6**
P value 95% CI OR P value 95% CI OR
95% CI
P value
OR
P value
OR
95% CI
Rebleeding Severe bleeding In-hospital complications Composite adverse outcomes
Table 4. Independent predictors of the composite adverse outcome, in-hospital complications, rebleeding and severe bleeding in patients with peptic ulcer disease-related upper gastrointestinal bleeding using multivariable logistic regression analysis
Journal of Digestive Diseases 2014; 15; 283–292
Antithrombotics on UGIB outcomes
289
patients, three died from uncontrolled UGIB or the interventions undertaken to control the bleeding, whereas the remaining 13 (81.3%) died from nonUGIB-related causes such as systemic malignancy (n = 4), multiorgan failure (n = 3), sepsis (n = 2) and cardiovascular and thromboembolic events (n = 4). Three patients died from myocardial infarction, including the one patient who was taking aspirin only. DISCUSSION Acute gastrointestinal bleeding has been increasingly reported to be related with the use of antithrombotics,7,8,19,20 and usually requires admission to hospital.21 Anticoagulants and aspirin may have an impact on the management and clinical outcomes of patients with UGIB.18,22 In this study, 59.8% of our patients with PUD-related UGIB were on antithrombotics and 35.2% were on aspirin only, which is in conformity with the recent studies (14.3–57%).1,12,14 We found that patients taking aspirin only had a lower morbidity and mortality and a shorter length of hospital stay than those not taking any antithrombotic agents. In contrast, those taking anticoagulants only had a higher frequency of adverse outcomes. According to our study, patients taking aspirin only had fewer adverse outcomes, even though they were elder than the overall population and had more comorbidities than those were not treated with any antithrombotics. Aging and comorbidities are known predictors of mortality and adverse clinical outcomes.3,5 Furthermore, the protective effect of aspirin may not be fully attributable to its known cardiovascular benefits, because most deaths are due to non-cardiovascular issues, as suggested by a placebocontrolled study.10 It is customary to discontinue aspirin in patients presenting with UGIB at our institution and elsewhere.9 Similar protective effects of aspirin against UGIBrelated deaths have been reported in studies that included both peptic and non-peptic ulcer patients.11,12 However, it is not clear whether the controls in these studies were taking other antithrombotics. Other studies, however, have reported no significant association between aspirin use and mortality1,14 although those taking non-steroidal anti-inflammatory drugs were included. Our study suggested that patients on aspirin who were admitted due to UGIB were less likely to have rebleeding, which might reflect an effect of aspirin
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
290
H Abu Daya et al.
discontinuation. However, this conclusion remains conflicting, and one study22 even associated a lowdose aspirin intake with longer length of hospital stay and increased requirement of blood transfusion. In contrast, patients treated with anticoagulants only had a higher incidence of adverse outcomes including severe hemorrhage and in-hospital complications, but not in-hospital mortality, which was probably due to the low mortality in our study. This is consistent with the observations that anticoagulants increase requirements for blood transfusion and are associated with a longer length of hospital stay.11,12,15–18 Marmo et al. showed that heparin, but not warfarin, increases mortality risk.13 Patients on heparin are usually hospitalized and develop in-hospital UGIB, which is a known risk factor for worse outcomes.13 Our observation that endoscopic therapy was predictive of adverse outcomes in patients on anticoagulants probably reflects the fact that these patients had stigmata of active or recent bleeding, which is associated with adverse outcomes or a higher rate of treatment failure in those taking anticoagulants. However, due to the small sample size on anticoagulants in our study, these results should be taken cautiously. In-hospital mortality in our study is within the reported range; in fact, it decreased from 3.9% to 2.7% from 1993 to 2006.2,5,14 Non-bleeding-related causes accounted for most deaths, which is in accordance with previous reports.5,23 The number of patients who died of acute renal failure in our study was much higher than that reported elsewhere (62.5% [10/16] vs 4.2%), as it was considered in the previous study to be an independent risk factor of death23 while we regarded it as part of multi-organ failure or the consequence of another major cause of death. Malignancies accounted for 30.8% of the non-bleeding-related deaths, which is in accordance with the previous results, while cardiovascular causes (30.8%) are higher than that reported (13–14%).5 We used rigorous criteria for the severe bleeding that occurred in 40.7% of the patients, which has been reported to be a risk factor for mortality.3,24,25 Our study, however, is the first to suggest that the intake of anticoagulants is an independent predictor of severe UGIB. The mechanism might be related to the impairment of hemostasis, which needs further investigation. Rebleeding is a predictor of mortality that occurred in 16.6% of our patients, the rate of which is similar to
Journal of Digestive Diseases 2014; 15; 283–292 previously reported rates.1,11 We showed that the risk of rebleeding was significantly decreased in patients taking aspirin but increased in those taking anticoagulants. Previous studies1,11,14–18,26 have suggested that the risk of rebleeding is not affected by antithrombotics. However, aspirin continuation seems to be associated with its increased risk of rebleeding.10 Aspirin use was a predictor for short length of hospital stay in our study, which is consistent with a better overall outcome for those patients. This is different from what has been reported, which might be due to varied study designs10 or the patients included in the study (both PUD and non-PUD-related UGIB).15 Another study22 suggested that patients having UGIB with aspirin use had a longer length of hospital stay, but the study included those with variceal bleeding. Our study has a number of strengths. First, data collection was performed using the ICD-9/10 codes that resulted in the identification of all the potential gastrointestinal bleeding cases, after which each case was reviewed individually using well-developed prespecified criteria. Furthermore, a comprehensive structured patient form for data collection was used. Second, we had a relatively large sample size with a homogenous cause of PUD-related UGIB. Third, we included a group of patients who were not taking any antiplatelet agents or anticoagulants. Moreover, this was a comprehensive examination of the association between aspirin or anticoagulant intake and a multitude of adverse outcomes in these patients. We recorded the etiologies of death in our patients in order to understand better the association between aspirin or anticoagulant use and patients’ death. Finally, the study was conducted in a tertiary care referral center where all the diagnostic and therapeutic procedures are standardized. There are some limitations in our study. Given that the study had a retrospective design and the fact that anticoagulant intake prior to the symptom presentation depended mainly on the patients’ recall, there might be a potential recall bias. However, we implemented a rigorous history-taking system at our institution where data on medication intake was gathered separately by the admitting physician as well as by the admitting nurse, and then both data sheets were compared for discrepancies. Moreover, these data were gathered from patients and their family members and compared to their most recent hospital or clinic visits, as most of our patients are usually regular attendants of our institutions. We also recog-
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 283–292 nized that observed differences in clinical outcomes might be due to a history of antithrombotic use or their discontinuation upon admission. Another limitation was the lack of complete data on Helicobacter pylori infection and endoscopic stigmata of bleeding, as these parameters were not recorded at the time of data collection. Therefore, we are unable to report the results at admission or during long-term followup. Because our study included patients over a long period of time, changes in clinical practice might have influenced the outcomes. However, to assess this temporal variation we divided the whole cohort into three consecutive 6-year time periods and repeated the same analysis; the results were similar in each time frame (data not shown). Finally, this study was performed at a single institution; therefore, its applicability to other populations requires further evaluation. In conclusion, our study suggested that aspirin intake might be associated with favorable clinical outcomes in PUD-related UGIB, while anticoagulants seem to be associated with worse outcomes. Further investigations are needed to validate these findings. Furthermore, if aspirin is found to improve outcomes in such patients, clinicians should be informed whether and when to resume aspirin in patients taking it for primary prophylaxis.
Antithrombotics on UGIB outcomes
6
7
8
9 10 11
12
13
14
ACKNOWLEDGEMENT Part of this research was presented as a poster presentation at the Digestive Disease Week 2012 held by the American Gastroenterological Association in San Diego, California, USA. REFERENCES 1 van Leerdam ME, Vreeburg EM, Rauws EA et al. Acute upper GI bleeding: did anything change? Time trend analysis of incidence and outcome of acute upper GI bleeding between 1993/1994 and 2000. Am J Gastroenterol 2003; 98: 1494–9. 2 Wang YR, Richter JE, Dempsey DT. Trends and outcomes of hospitalizations for peptic ulcer disease in the United States, 1993 to 2006. Ann Surg 2010; 251: 51–8. 3 Marmo R, Koch M, Cipolletta L et al. Predictive factors of mortality from nonvariceal upper gastrointestinal hemorrhage: a multicenter study. Am J Gastroenterol 2008; 103: 1639–47; quiz 1648. 4 Hearnshaw SA, Logan RF, Lowe D, Travis SP, Murphy MF, Palmer KR. Acute upper gastrointestinal bleeding in the UK: Patient characteristics, diagnoses and outcomes in the 2007 UK audit. Gut 2011; 60: 1327–35. 5 Sung JJ, Tsoi KK, Ma TK, Yung MY, Lau JY, Chiu PW. Causes of mortality in patients with peptic ulcer bleeding: a
15
16 17
18
19
20
291
prospective cohort study of 10,428 cases. Am J Gastroenterol 2010; 105: 84–9. Cooper GS, Yuan Z, Rosenthal GE, Chak A, Rimm AA. Lack of gender and racial differences in surgery and mortality in hospitalized Medicare beneficiaries with bleeding peptic ulcer. J Gen Intern Med 1997; 12: 485–90. Theocharis GJ, Thomopoulos KC, Sakellaropoulos G, Katsakoulis E, Nikolopoulou V. Changing trends in the epidemiology and clinical outcome of acute upper gastrointestinal bleeding in a defined geographical area in Greece. J Clin Gastroenterol 2008; 42: 128–33. García Rodríguez LA, Lin KJ, Hernández-Díaz S, Johansson S. Risk of upper gastrointestinal bleeding with low-dose acetylsalicylic acid alone and in combination with clopidogrel and other medications. Circulation 2011; 123: 1108–15. Sostres C, Lanas A. Should prophylactic low-dose aspirin therapy be continued in peptic ulcer bleeding? Drugs 2011; 71: 1–10. Sung JJ, Lau JY, Ching JY et al. Continuation of low-dose aspirin therapy in peptic ulcer bleeding: a randomized trial. Ann Intern Med 2010; 152: 1–9. Lanas A, Aabakken L, Fonseca J et al. Clinical predictors of poor outcomes among patients with nonvariceal upper gastrointestinal bleeding in Europe. Aliment Pharmacol Ther 2011; 33: 1225–33. Åhsberg K, Höglund P, Staël von Holstein C. Mortality from peptic ulcer bleeding: the impact of comorbidity and the use of drugs that promote bleeding. Aliment Pharmacol Ther 2010; 32: 801–10. Marmo R, Koch M, Cipolletta L et al.; Italian registry on upper gastrointestinal bleeding (Progetto Nazionale Emorragie Digestive–PNED 2). Predicting mortality in non-variceal upper gastrointestinal bleeders: validation of the Italian PNED score and prospective comparison with the Rockall score. Am J Gastroenterol 2010; 105: 1284–91. Mose H, Larsen M, Riis A, Johnsen SP, Thomsen RW, Sørensen HT. Thirty-day mortality after peptic ulcer bleeding in hospitalized patients receiving low-dose aspirin at time of admission. Am J Geriatr Pharmacother 2006; 4: 244–50. Ortiz V, Ortuño J, Rodríguez-Soler M, Iborra M, Garrigues V, Ponce J. Outcome of non-variceal acute upper gastrointestinal bleeding in patients with antithrombotic therapy. Digestion 2009; 80: 89–94. Choudari CP, Rajgopal C, Palmer KR. Acute gastrointestinal haemorrhage in anticoagulated patients: diagnoses and response to endoscopic treatment. Gut 1994; 35: 464–6. Thomopoulos KC, Mimidis KP, Theocharis GJ, Gatopoulou AG, Kartalis GN, Nikolopoulou VN. Acute upper gastrointestinal bleeding in patients on long-term oral anticoagulation therapy: endoscopic findings, clinical management and outcome. World J Gastroenterol 2005; 11: 1365–8. Wolf AT, Wasan SK, Saltzman JR. Impact of anticoagulation on rebleeding following endoscopic therapy for nonvariceal upper gastrointestinal hemorrhage. Am J Gastroenterol 2007; 102: 290–6. Barada K, Abdul-Baki H, El Hajj II, Hashash JG, Green PH. Gastrointestinal bleeding in the setting of anticoagulation and antiplatelet therapy. J Clin Gastroenterol 2009; 43: 5–12. Hallas J, Dall M, Andries A et al. Use of single and combined antithrombotic therapy and risk of serious upper gastrointestinal bleeding: population based case-control study. BMJ 2006; 333: 726.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
292
H Abu Daya et al.
21 Pirmohamed M, James S, Meakin S et al. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ 2004; 329: 15–9. 22 Taha AS, Angerson WJ, Knill-Jones RP, Blatchford O. Clinical outcome in upper gastrointestinal bleeding complicating low-dose aspirin and antithrombotic drugs. Aliment Pharmacol Ther 2006; 24: 633–6. 23 Wong GL, Wong VW, Chan Y et al. High incidence of mortality and recurrent bleeding in patients with Helicobacter pylori-negative idiopathic bleeding ulcers. Gastroenterology 2009; 137: 525–31. 24 Chiu PW, Ng EK, Cheung FK et al. Predicting mortality in patients with bleeding peptic ulcers after therapeutic
Journal of Digestive Diseases 2014; 15; 283–292 endoscopy. Clin Gastroenterol Hepatol 2009; 7: 311–6; quiz 253. 25 Marmo R, Del Piano M, Rotondano G et al. Mortality from nonulcer bleeding is similar to that of ulcer bleeding in high-risk patients with nonvariceal hemorrhage: a prospective database study in Italy. Gastrointest Endosc 2012; 75: 263–72. 26 Rubin TA, Murdoch M, Nelson DB. Acute GI bleeding in the setting of supratherapeutic international normalized ratio in patients taking warfarin: endoscopic diagnosis, clinical management, and outcomes. Gastrointest Endosc 2003; 58: 369–73.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 283–292
doi: 10.1111/1751-2980.12140
Original article
Opposing effects of aspirin and anticoagulants on morbidity and mortality in patients with upper gastrointestinal bleeding Hussein ABU DAYA,*,† Mohamad ELOUBEIDI,† Hani TAMIM,‡,§ Houssam HALAWI,‡ Ahmad H. MALLI,‡ Don C. ROCKEY¶ & Kassem BARADA†
¶
*Department of Internal Medicine, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, Department of Internal Medicine, Medical University of South Carolina, Charleston, South Carolina, USA; and †Division of Gastroenterology and Hepatology, ‡Department of Internal Medicine, and §Biostatistics Support Unit, Clinical Research Institute, American University of Beirut Medical Center, Beirut, Lebanon
OBJECTIVE: We aimed to determine the effect of antithrombotics on in-hospital mortality and morbidity in patients with peptic ulcer disease-related upper gastrointestinal bleeding (PUD-related UGIB). METHODS: The study cohort was retrospectively selected from a tertiary center database of patients with PUD-related UGIB, defined as bleeding due to gastric or duodenal ulcers, or erosive duodenitis, gastritis or esophagitis. Outcomes were compared among patient groups based on their antithrombotic medications before admission. Patients on no antithrombotics served as controls. The composite adverse outcomes, in-hospital mortality, rebleeding and/or need for surgery were measured. Severe bleeding and in-hospital complications were also recorded. RESULTS: Of 398 patients with PUD-related UGIB, 44.5% were on aspirin or anticoagulants only. The
composite adverse outcome was most common in patients taking anticoagulants only (40.5%), intermediate in controls (23.1%) and least in those taking aspirin only (12.1%). On multivariate analysis, patients taking aspirin alone had a significantly lower risk of adverse outcome events (odds ratio [OR] 0.4, 95% CI 0.2–0.8) and a shorter length of hospital stay (regression coefficient = −3.4, 95% CI [−6.6, −0.6]). In contrast, taking anticoagulants was associated with a greater risk of adverse outcome events (OR 2.3, 95% CI 1.0–5.3), severe bleeding (OR 2.6, 95% CI 1.2–5.8) and in-hospital complications (OR 2.9, 95% CI 1.3– 6.6). CONCLUSIONS: Patients with PUB-related UGIB while taking aspirin had fewer adverse outcomes compared with those taking anticoagulants. Aspirin may have beneficial effects in this population.
KEY WORDS: anticoagulant, gastrointestinal hemorrhage, morbidity, peptic ulcer hemorrhage, platelet aggregation inhibitor.
Correspondence to: Kassem BARADA, Division of Gastroenterology and Hepatology, American University of Beirut Medical Center, Riad El Solh, Beirut 1107 2020, Lebanon. Email: [email protected] Conflicts of interest: None. © 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
INTRODUCTION Upper gastrointestinal bleeding (UGIB) is associated with high morbidity and mortality.1 In the USA the annual estimated hospitalization nationwide due to peptic ulcer disease (PUD) ranged from 156 108 to 222 601, with a decreasing trend from 1993 to 2006.2 Up to two-thirds of UGIB admissions are due to PUD.3
283
284
H Abu Daya et al.
Despite the development of treatment modalities, the 30-day mortality in patients suffering from UGIB was reported to be 8.9% in the UK,4 6.2% in Hong Kong SAR, China and 8.5% in the USA.5,6 Antiplatelet agents and anticoagulants are increasingly applied in the clinic setting for the treatment and prophylaxis of cardiovascular diseases, especially in elderly populations.7 Monotherapy with antiplatelet agents or anticoagulants is associated with an increasing risk of UGIB8 and aspirin has been verified to play an important role in the pathogenesis of PUD. Therefore, it is common practice to discontinue these medications in patients with UGIB for days, even weeks, during and after bleeding episode.9 Although the continuation of low-dose aspirin will increase the risk of rebleeding, it has been reported to be able to reduce the all-cause mortality rate (1.3% vs 12.9% compared with placebo) in a small sample of patients.10 Current evidence on the efficacy of aspirin or anticoagulants on the clinical outcomes of patients with UGIB is controversial, as some authors10–13 have reported that aspirin decreases the mortality of patients while others1,14,15 have suggested these drugs have no effects. Similarly, conflicting results on the effects of anticoagulants such as warfarin have been reported.11,12,16–18 These drugs have been reported to increase the patients’ mortality significantly13 or having no influence on the disease course of UGIB, except that they are associated with the patients’ prolonged hospital stay.15 In this study we aimed to determine the impact of aspirin and anticoagulants on the clinical outcomes of patients who were admitted to hospital and hospitalized due to PUD-related UGIB, and to investigate the etiology of death in patients who had a fatal outcome. PATIENTS AND METHODS Patients Patients who were admitted to hospital and hospitalized in the Division of Gastroenterology and Hepatology, American University of Beirut Medical Center (Beirut, Lebanon) from 1993 to 2010 with evidence of gross UGIB were identified using the International Statistical Classification of Diseases and Related Health Problems (ICD)-9/ICD-10 coding system. UGIB was defined as having the symptoms and signs of reported or witnessed hematemesis, melena or hematochezia with an identified source of
Journal of Digestive Diseases 2014; 15; 283–292 bleeding in the upper gastrointestinal tract by esophagogastroduodenoscopy (EGD). PUD was considered to be present in those with gastric or duodenal ulcers or esophageal, gastric or duodenal erosions. All patients diagnosed with PUD-related UGIB during the study period were recruited in the study. Exclusion criteria were: (i) non-PUD-related UGIB; (ii) small intestinal or lower gastrointestinal bleeding, defined as hematochezia or melena with the bleeding source identified by colonoscopy or small intestinal followthrough, push enteroscopy, ileoscopy or computed tomography examination of the abdomen and pelvis, respectively, but in the absence of UGIB lesion on EGD; and (iii) occult UGIB. The study was approved by the Institutional Review Board. Antithrombotic medications A history of antithrombotic medications used at the time of diagnosis was obtained by reviewing the medical records of the patients as well as the notes from the nurses and physicians that contained sections devoted to the patients’ intake of medications at home. The patients were then divided into five groups based on their drug intake before admission: (i) those who had not taken any anticoagulants or antiplatelet agents (group A); (ii) those taking anticoagulants (warfarin, heparin or low-molecularweight heparin) only (group B); (iii) those taking aspirin only (group C); (iv) those taking any antiplatelet agent except aspirin (clopidogrel, ticlopidine or dipyridamole) with or without aspirin (group D); and (v) those taking antiplatelet agents combined with anticoagulants (group E). Study variables Patients’ characteristics including their individual variables such as age and gender, past medical history, clinical manifestations, vital signs and blood test results when they had been taken to the Emergency Room, their management in the Emergency Room and in hospital, diagnostic EGD findings, types of therapeutic endoscopic procedures, angiography and embolization, if any, and surgical treatment, if any, as well as their in-hospital mortality. The etiologies of in-hospital mortality were documented and divided into the following types: (i) cardiovascular/ thromboembolic events such as myocardial infarction, pulmonary embolus and stroke; (ii) a defined focus of infection associated with septic shock; (iii) death due to uncontrolled gastrointestinal bleeding or
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 283–292
Antithrombotics on UGIB outcomes
complications related to the inventions used to control bleeding; (iv) systemic cancer; and (v) multiorgan failure. Definition of clinical outcomes Primary or composite adverse outcomes were defined as the occurrence of at least one of the following factors: in-hospital mortality, need for surgery (requiring any kinds of surgical procedures to control bleeding) and rebleeding. Secondary outcomes were defined as severe bleeding, in-hospital complications and the length of hospital stay. The need for and the amounts of blood transfusion given were recorded. Rebleeding was defined as the recurrence of hematemesis, coffee ground emesis or melena 24 h after the initial endoscopic evaluation or hemostatic therapy and initial stabilization, accompanied by either a change in vital signs or a decrease in hemoglobin level by at least 20 g/L. Rebleeding events during the same hospitalization or requiring another hospitalization were combined upon the retrieval of data in patients with ICD-9/10 codes of UGIB. In the case of multiple admissions, only the first admission was included and subsequent re-admissions were considered as rebleeding episodes if they took place within one month of the initial admission. Severe bleeding was defined based on the following criteria: (i) hypotension: systolic blood pressure < 90 mmHg at admission; (ii) tachycardia: >119 beats/min at admission; or (iii) transfusion of more than 750 mL packed red blood cells. In-hospital complications included cardiovascular or thromboembolic events (myocardial infarction, angina, deep vein thrombosis, pulmonary embolism, stroke and transient ischemic attack, etc.), infectious complications including pneumonia, urinary tract infection, skin infections, sepsis and other complications such as acute respiratory distress syndrome (ARDS), renal failure, the need for mechanical ventilation and disseminated intravascular coagulation. Statistical analysis Statistical analyses were performed using SAS 9.1 (SAS Institute, Cary, NC, USA). Categorical variables were expressed as numbers and percentages, and continuous data were expressed as mean ± standard deviation. Bivariate analyses were performed using the χ2 test or Student’s t-test, where appropriate. To control for the effect of potentially confounding variables multivariate analyses were carried out. For
285
Gastrointestinal bleeding (n = 1077)
LGIB (n = 220)
NV-UGIB (n = 717)
Non-PUD-related UGIB‡ (n = 319)
Admissions excluded† (n = 140)
PUD-related UGIB (n = 398)
Figure 1. Flowchart of patients’ enrollment and grouping in patients with peptic ulcer disease (PUD)-related upper gastrointestinal bleeding (UGIB). †Esophageal variceal bleeding (n = 85), small intestinal bleeding (n = 9) and occult gastrointestinal bleeding (n = 46). ‡Causes of nonPUD-related include Mallory–Weiss tear, arteriovenous malformation, Dieulafoy’s lesion, malignancy and hiatal hernia. LGIB, lower gastrointestinal bleeding; NV, non-variceal.
categorical outcomes (such as the composite adverse outcomes), a multivariate logistic regression was performed and the odds ratio (OR) and 95% confidence interval (CI) were calculated. On the other hand, for continuous outcomes (such as the length of hospital stay), multivariate linear regression was carried out where the β coefficient and 95% CI were recorded. Variables included in the regression model were those of either statistical or clinical significance. P ≤ 0.05 was considered statistically significant. RESULTS Characteristics of the patients Altogether 1077 patients with acute gastrointestinal bleeding were diagnosed and treated in our hospital from 1993 to 2010. Of these patients, 679 were excluded from the study for lower gastrointestinal bleeding (n = 220), esophageal variceal bleeding (n = 85), small intestinal bleeding (n = 9), occult gastrointestinal bleeding (n = 46) and non-PUD-related UGIB (n = 319). Finally, 398 patients aged 65.1 ± 15.2 years were enrolled in the study (Fig. 1), with a significant male predominance (male vs female: 73.9% vs 26.1%; Table 1). Patients with PUD-related UGIB in the group A were significantly younger and had less comorbidity than those taking antithrombotics (P < 0.01). The proportion of patients using proton pump inhibitor (PPI) before presentation was similar among the different groups (8% in group A vs 7.1% in group C [P = 0.80] and 16.2% in group B [P = 0.10], respectively). The frequency of patients having alcohol intake was significantly lower in group B than in group
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
11 (29.7) 16 (43.2) 25 (67.6) 17 (45.9)***
51 (31.9) 72 (45.0) 115 (71.9) 83 (51.9)
5 (13.5) 17 (45.9) 11 (29.7) 2 (5.4) 4 (10.8)
35 (21.9) 81 (50.6) 29 (18.1) 4 (2.5) 26 (16.3) 0.24 ± 0.06 2.9 ± 1.8**
12 (32.4)** 26 (70.3)** 18 (48.6)** 10 (27.0)** 7 (18.9) 7 (18.9)
21 (13.1) 50 (31.3) 19 (11.9) 2 (1.3) 29 (18.1) 39 (24.4)
0.27 ± 0.08 1.2 ± 0.7
71.1 ± 11.0** 21 (56.8)
60.5 ± 18.3 112 (70.0)
Group B (n = 37)
55 (39.3) 60 (42.9) 105 (75.0) 88 (62.9)***
0.27 ± 0.06 1.0 ± 0.2*
21 (15.0) 77 (55.0) 27 (19.3) 5 (3.6) 28 (20.0)
45 (32.1)** 76 (54.3)** 50 (35.7)** 4 (2.9) 13 (9.3)* 23 (16.4)
66.9 ± 12.0** 110 (78.6)
Group C (n = 140)
13 (38.2) 14 (41.2) 25 (73.5) 20 (58.8)
0.26 ± 0.04 1.2 ± 0.6
7 (20.6) 8 (23.5)** 12 (35.3)* 4 (11.8)** 11 (32.4)*
11 (32.4)** 25 (73.5)** 26 (76.5)** 0 (0) 5 (14.7) 3 (8.8)*
70.0 ± 13.0** 28 (82.4)
Group D (n = 34)
8 (29.6) 8 (29.6) 15 (55.6) 17 (63.0)
0.26 ± 0.07 2.2 ± 2.0**
4 (14.8) 16 (59.3) 6 (22.2) 2 (7.4) 5 (18.5)
9 (33.3)** 16 (59.3)** 21 (77.8)** 6 (22.2)** 2 (7.4) 0 (0)**
70.0 ± 9.0** 23 (85.2)
Group E (n = 27)
138 (34.7) 170 (42.7) 285 (71.6) 225 (56.5)
0.27 ± 0.07 1.5 ± 1.1
72 (18.1) 199 (50.0) 85 (21.4) 17 (4.3) 74 (18.6)
98 (24.6) 193 (48.5) 134 (33.7) 22 (5.5) 56 (14.1) 72 (18.1)
65.1 ± 15.2 294 (73.9)
Total (n = 398)
H Abu Daya et al.
*P < 0.05, **P < 0.01, ***P = 0.054 compared with group A. †Each symptom and endoscopic diagnosis (including patients with more than one symptom) are included. ‡Patients with gastric ulcer and/or duodenal ulcer. Univariate analysis is performed with group A (those who are not taking any antithrombotics) as the control group. CAD, coronary artery disease; Group A, patients who have taken no anticoagulants or antiplatelet agents; group B, patients taking anticoagulants (warfarin, heparin or low-molecular-weight heparin) only; group C, patients taking aspirin only (group C); group D, patients taking any antiplatelet agents except aspirin (clopidogrel, ticlopidine or dipyridamole) with or without aspirin; group E, patients taking antiplatelet agents combined with anticoagulants; INR, international normalized ratio; NSAID, non-steroidal anti-inflammatory drugs; PMH, past medical history; SD, standard deviation.
Age, years (mean ± SD) Male gender, n (%) PMH, n (%) Diabetes Hypertension CAD Atrial fibrillation Cancer Non-aspirin NSAIDs use Symptoms at presentation, n (%)† Hematemesis Melena Hematemesis + melena Hematochezia Syncope Initial laboratory results Hematocrit (mean ± SD) INR (mean ± SD) Endoscopic diagnosis, n (%)† Gastric ulcer Duodenal ulcer Peptic ulcer‡ Erosion
Group A (n = 160)
Table 1. Individual and clinical characteristics of patients with peptic ulcer disease-related upper gastrointestinal bleeding, comparing those taking different antiplatelet agents and anticoagulants
286
Journal of Digestive Diseases 2014; 15; 283–292
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
287
*P < 0.05; **P < 0.01, and ***P = 0.056 compared with the group A. †Composite adverse outcomes are defined as the occurrence of in-hospital mortality, need for surgery or rebleeding. Group A, patients have not taken any anticoagulants or antiplatelet agents; group B, those taking anticoagulants (warfarin, heparin or low-molecular-weight heparin) only; group C, those taking aspirin only (group C); group D, those taking any antiplatelet agents except aspirin (clopidogrel, ticlopidine or dipyridamole) with or without aspirin; and group E, those taking antiplatelet agents combined with anticoagulants.
16 (4.0) 24 (6.0) 66 (16.6) 79 (19.8) 162 (40.7) 136 (34.2) 96 (24.1) 301 (75.6) 1250 ± 2500 7±9 0 (0) 0 (0) 4 (14.8) 4 (14.8) 16 (59.3)* 15 (55.6)** 7 (25.9) 23 (85.2)* 1250 ± 1250 10 ± 11
Total (N = 398) Group E (n = 27) Group D (n = 34)
1 (2.9) 0 (0) 6 (17.6) 6 (17.6) 9 (26.5) 10 (29.4) 11 (32.4) 30 (88.2)** 750 ± 500 5±4 1 (0.7)** 4 (2.9)* 14 (10.0)* 17 (12.1)* 53 (37.9) 39 (27.9) 35 (25.0) 109 (77.9)* 875 ± 750* 5 ± 4* 4 (10.8) 5 (13.5) 12 (32.4)*** 15 (40.5)* 24 (64.9)** 24 (64.9)** 8 (21.6) 35 (94.6)** 2750 ± 6500* 12 ± 15***
Mortality (0.7% vs 6.3%, P < 0.01) and rebleeding rate (10.0% vs 18.8%, P < 0.05) were less frequent in the patients treated with aspirin only (group C) than those taking neither antiplatelet agents nor anticoagulants (group A), while those taking anticoagulants only (group B) had both the highest mortality and the highest rebleeding rate among all the five groups (10.8% and 32.4%, respectively) (Table 2). In addition, the length of hospital stay was significantly shorter in patients in group C than in groups A, B and E, but was similar to that in group D. A high proportion of patients in group C had a blood transfusion; however, they were given fewer volumes of blood transfusion per patient than group A (875 ± 750 mL vs 1250 ± 1750 mL, P < 0.05). In contrast, patients in group B were most likely to require a blood transfusion and received the highest volumes of blood transfusions, and they had the longest length of hospital stay among all the groups, although this was of borderline significance (P = 0.056). Similar overall associations were observed in patients in group D compared with the group A, as they had lower mortality (2.9% vs 6.3%), less need for surgery (0% vs 9.4%) and composite adverse outcome rates (17.6% vs 23.1%) together with a shorter duration of hospital
10 (6.3) 15 (9.4) 30 (18.8) 37 (23.1) 60 (37.5) 48 (30.0) 35 (21.9) 104 (65.0) 1250 ± 1750 8 ± 11
Clinical outcomes of the patients
Mortality n (%) Surgery n (%) Rebleeding n (%) Composite adverse outcomes† n (%) Severe bleeding n (%) In-hospital complications n (%) Endoscopic therapy n (%) Blood transfusion n (%) Blood transfusion, mL (mean ± SD) Length of hospital stay, days (mean ± SD)
In total, 285 (71.6%) patients were found to have peptic ulcers (gastric and/or duodenal ulcers) and 225 (56.5%) had erosions. More than 20% of the patients had multiple lesions. There was no difference in the frequency of duodenal or gastric ulcers among the groups.
Group C (n = 140)
Patients in the cohort had evidence of substantial blood loss, with a mean hematocrit level of 0.27 ± 0.07 (Table 1). Melena was the most common presentation, occurring in half the patients, followed by hematemesis combined with melena (21.4%), syncope (18.6%), hematemesis only (18.1%) and hematochezia (4.3%), respectively. The mean international normalized ratio (INR) was higher in patients who were using anticoagulants only (2.9 ± 1.8, P < 0.01) but lower in those using aspirin only (1.0 ± 0.2, P < 0.05), respectively, than those in group A who were treated with neither anticoagulants nor antiplatelet agents (1.2 ± 0.7).
Group B (n = 37)
Clinical presentations and the etiology of UGIB
Group A (n = 160)
A (18.6% vs 3.3%, P = 0.04), but it was otherwise similar in the different groups.
Antithrombotics on UGIB outcomes Table 2. Frequencies of clinical outcomes in patients with peptic ulcer disease-related upper gastrointestinal bleeding and comparison between patients taking any anticoagulants (AC) or antiplatelet (AP) and controls
Journal of Digestive Diseases 2014; 15; 283–292
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
0.0001 0.7000 0.0001 0.0001
56 (34.6) 53 (32.7) 36 (22.2) 17 (10.5) NA 53 (46.9) 24 (28.6) 65 ± 15 115 (43.9) 90 (34.3) 46 (17.6) 11 (4.2) 84 (32.1) 101 (47.4) 13 (10.4) 62 ± 15 NA 0.01 0.03 0.94
46 (33.8) 44 (32.4) 30 (22.0) 16 (11.8) 78 (57.4) 39 (44.8) 24 (33.3) 71 ± 13 128 (40.1) 111 (34.8) 58 (18.2) 22 (6.9) NA 123 (50.0) 22 (15.2) 65 ± 15
0.71
0.004 0.010 0.330 0.940
33 (41.8) 23 (29.1) 18 (22.8) 5 (6.3) NA† 17 (31.5) 15 (28.8) 65 ± 17 8 (50.0) 6 (37.4) 1 (6.3) 1 (6.3) 12 (75.0) 1 (9.1) 4 (28.6) 65 ± 17
153 (40.1) 128 (33.5) 75 (19.6) 26 (6.8) 150 (39.3) 139 (48.1) 33 (18.0) 65 ± 15
0.590
†Composite adverse outcomes (primary outcomes) include in-hospital mortality, rebleeding or need for surgery. ‡Comorbidity includes coronary artery disease, congestive heart failure, cerebral vascular accident, systemic cancer, chronic renal failure and diabetes mellitus. §Percentages calculated as: (patients with the variables [yes or no] in group C or B)/(patients with the variables [yes or no] in group C or B + those with the variables [yes or no] in group A). NA, not applicable.
NA 0.900 0.002 0.990
26 (39.4) 18 (27.3) 17 (25.7) 5 (7.6) NA 14 (31.8) 12 (28.6) 67 ± 15 105 (44.5) 81 (34.3) 40 (16.9) 10 (4.3) NA 87 (46.5) 13 (11.5) 65 ± 15
0.02 0.01
Yes P value No Yes P value No Yes P value No
NA 0.03 0.07 0.30 135 (40.7) 116 (34.9) 59 (17.8) 22 (6.6) NA 126 (49.2) 25 (16.1) 65 ± 15
0.40
P value No No
P value
Yes
Rebleeding Severe bleeding
Number of comorbidities, n (%)‡ 0 1 2 ≥3 Severe bleeding, n (%) Aspirin only, n (%)§ Anticoagulants only, n (%)§ Age, years (mean ± SD)
To further explore the protective effect of aspirin against in-hospital mortality, we determined the cause of death in the 16 patients, in which one was using aspirin only, one used a combination of aspirin and clopidogrel, three used heparin, one used coumadin and ten took no antithrombotics. Among these
Yes
On multivariate analysis, elder age (> 60 years) was found to be an independent predictor of in-hospital complications (OR 3.2, 95% CI 1.6–6.1, P = 0.001; Table 4). In addition, intake of aspirin alone at presentation was independently associated with fewer adverse outcomes (OR 0.4, 95% CI 0.2–0.8, P = 0.01) and shorter length hospital stay (r = −3.4, 95% CI [−6.6, −0.6], P = 0.02). On the other hand, being on anticoagulants was associated with a greater risk of adverse outcomes, in-hospital complications and severe bleeding (P = 0.04, 0.007 and 0.02, respectively). Finally, endoscopic therapy was an independent predictor of rebleeding (P = 0.01) and adverse outcomes (P = 0.01) in patients on anticoagulants. PPI use was not associated with the composite adverse outcomes (OR 0.7, 95% CI 0.2–2.4, P = 0.50), in-hospital complications (OR 0.9, 95% CI 0.3–2.3, P = 0.80), severe bleeding (OR 0.5, 95% CI 0.2–1.3, P = 0.20) or rebleeding (OR 0.5, 95% CI 0.1–2.3, P = 0.40).
In-hospital complications
Patients who had in-hospital complications were on average a decade elder than those who did not (P = 0.0001) (Table 3). Similarly, the frequencies of patients with in-hospital complications and severe bleeding increased as the number of their comorbidities increased (P = 0.01 and 0.02, respectively). Moreover, the frequency of severe bleeding was higher in patients who suffered in-hospital death and complications than in those who did not (P = 0.004 and 0.0001, respectively). Finally, patients who died or had composite adverse outcomes were less likely to be on aspirin only at presentation than those who did not (9.1% vs 48.1%, P = 0.010; and 31.5% vs 50.0%, P = 0.01). In contrast, those had composite adverse outcomes (P = 0.03), in-hospital complications (P = 0.0001) and severe bleeding (P = 0.002) were more likely to be on anticoagulants only than those who did not; however, those on anticoagulants were elder and had more comorbidities than group A (1.30 ± 1.00 vs 0.56 ± 0.81, P < 0.0001).
Composite adverse outcome†
stay (5 ± 4 days vs 8 ± 11 days). Interestingly, there were no deaths in patients taking both antiplatelet agents and anticoagulants (group E) although they were similar in severity of bleeding and rates of in-hospital complications to group B.
Journal of Digestive Diseases 2014; 15; 283–292
Mortality
H Abu Daya et al.
Table 3. Predictors of in-hospital mortality, composite adverse outcomes, rebleeding, severe bleeding and in-hospital complications using univariate analysis in patients with peptic ulcer disease-related upper gastrointestinal bleeding
288
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Anticoagulatn use
*P < 0.05, **P ≤ 0.01. †Multivariable analysis for aspirin use and AC use is performed with group A (those who were not taking any antithrombotics) as the control group. ‡Comorbidities include coronary artery disease, congestive heart failure, cerebral vascular accident, systemic cancer, chronic renal failure and diabetes mellitus. §Age ≤ 60 years vs >60 years.
0.01 0.80 0.07 0.06 0.14 0.9 0.5 0.01 0.2–0.8 0.6–1.4 0.9–4.6 0.9–3.9 0.8–4.6 0.7–1.6 0.6–3.2 1.2–5.5 0.4–1.2 0.9–1.7 1.6–6.1 0.6–2.1 1.3–6.6 0.9–1.8 1.0–4.5 0.5–2.2 Aspirin use
Aspirin only† Comorbidities‡ Age§ Endoscopic therapy Anticoagulants only† Comorbidities‡ Age § Endoscopic therapy
0.4** 0.9 1.3 1.7 2.3* 0.9 1.2 2.6**
0.2–0.8 0.6–1.4 0.6–2.5 0.8–3.2 1.0–5.3 0.6–1.4 0.6–2.6 1.3–5.4
0.01 0.69 0.47 0.14 0.04 0.70 0.60 0.01
0.7 1.2 3.2** 1.2 2.9** 1.3 2.1* 1.0
0.15 0.20 0.001 0.60 0.007 0.20 0.04 0.90
0.9 1.2 1.1 1.5 2.6* 1.3 1.0 1.3
0.6–1.5 0.9–1.6 0.7–1.9 0.9–2.7 1.2–5.8 0.9–1.9 0.5–1.9 0.6–2.5
0.70 0.20 0.70 0.10 0.02 0.10 0.90 0.50
0.4** 1.0 2.1 1.9 1.9 1.0 1.4 2.6**
P value 95% CI OR P value 95% CI OR
95% CI
P value
OR
P value
OR
95% CI
Rebleeding Severe bleeding In-hospital complications Composite adverse outcomes
Table 4. Independent predictors of the composite adverse outcome, in-hospital complications, rebleeding and severe bleeding in patients with peptic ulcer disease-related upper gastrointestinal bleeding using multivariable logistic regression analysis
Journal of Digestive Diseases 2014; 15; 283–292
Antithrombotics on UGIB outcomes
289
patients, three died from uncontrolled UGIB or the interventions undertaken to control the bleeding, whereas the remaining 13 (81.3%) died from nonUGIB-related causes such as systemic malignancy (n = 4), multiorgan failure (n = 3), sepsis (n = 2) and cardiovascular and thromboembolic events (n = 4). Three patients died from myocardial infarction, including the one patient who was taking aspirin only. DISCUSSION Acute gastrointestinal bleeding has been increasingly reported to be related with the use of antithrombotics,7,8,19,20 and usually requires admission to hospital.21 Anticoagulants and aspirin may have an impact on the management and clinical outcomes of patients with UGIB.18,22 In this study, 59.8% of our patients with PUD-related UGIB were on antithrombotics and 35.2% were on aspirin only, which is in conformity with the recent studies (14.3–57%).1,12,14 We found that patients taking aspirin only had a lower morbidity and mortality and a shorter length of hospital stay than those not taking any antithrombotic agents. In contrast, those taking anticoagulants only had a higher frequency of adverse outcomes. According to our study, patients taking aspirin only had fewer adverse outcomes, even though they were elder than the overall population and had more comorbidities than those were not treated with any antithrombotics. Aging and comorbidities are known predictors of mortality and adverse clinical outcomes.3,5 Furthermore, the protective effect of aspirin may not be fully attributable to its known cardiovascular benefits, because most deaths are due to non-cardiovascular issues, as suggested by a placebocontrolled study.10 It is customary to discontinue aspirin in patients presenting with UGIB at our institution and elsewhere.9 Similar protective effects of aspirin against UGIBrelated deaths have been reported in studies that included both peptic and non-peptic ulcer patients.11,12 However, it is not clear whether the controls in these studies were taking other antithrombotics. Other studies, however, have reported no significant association between aspirin use and mortality1,14 although those taking non-steroidal anti-inflammatory drugs were included. Our study suggested that patients on aspirin who were admitted due to UGIB were less likely to have rebleeding, which might reflect an effect of aspirin
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
290
H Abu Daya et al.
discontinuation. However, this conclusion remains conflicting, and one study22 even associated a lowdose aspirin intake with longer length of hospital stay and increased requirement of blood transfusion. In contrast, patients treated with anticoagulants only had a higher incidence of adverse outcomes including severe hemorrhage and in-hospital complications, but not in-hospital mortality, which was probably due to the low mortality in our study. This is consistent with the observations that anticoagulants increase requirements for blood transfusion and are associated with a longer length of hospital stay.11,12,15–18 Marmo et al. showed that heparin, but not warfarin, increases mortality risk.13 Patients on heparin are usually hospitalized and develop in-hospital UGIB, which is a known risk factor for worse outcomes.13 Our observation that endoscopic therapy was predictive of adverse outcomes in patients on anticoagulants probably reflects the fact that these patients had stigmata of active or recent bleeding, which is associated with adverse outcomes or a higher rate of treatment failure in those taking anticoagulants. However, due to the small sample size on anticoagulants in our study, these results should be taken cautiously. In-hospital mortality in our study is within the reported range; in fact, it decreased from 3.9% to 2.7% from 1993 to 2006.2,5,14 Non-bleeding-related causes accounted for most deaths, which is in accordance with previous reports.5,23 The number of patients who died of acute renal failure in our study was much higher than that reported elsewhere (62.5% [10/16] vs 4.2%), as it was considered in the previous study to be an independent risk factor of death23 while we regarded it as part of multi-organ failure or the consequence of another major cause of death. Malignancies accounted for 30.8% of the non-bleeding-related deaths, which is in accordance with the previous results, while cardiovascular causes (30.8%) are higher than that reported (13–14%).5 We used rigorous criteria for the severe bleeding that occurred in 40.7% of the patients, which has been reported to be a risk factor for mortality.3,24,25 Our study, however, is the first to suggest that the intake of anticoagulants is an independent predictor of severe UGIB. The mechanism might be related to the impairment of hemostasis, which needs further investigation. Rebleeding is a predictor of mortality that occurred in 16.6% of our patients, the rate of which is similar to
Journal of Digestive Diseases 2014; 15; 283–292 previously reported rates.1,11 We showed that the risk of rebleeding was significantly decreased in patients taking aspirin but increased in those taking anticoagulants. Previous studies1,11,14–18,26 have suggested that the risk of rebleeding is not affected by antithrombotics. However, aspirin continuation seems to be associated with its increased risk of rebleeding.10 Aspirin use was a predictor for short length of hospital stay in our study, which is consistent with a better overall outcome for those patients. This is different from what has been reported, which might be due to varied study designs10 or the patients included in the study (both PUD and non-PUD-related UGIB).15 Another study22 suggested that patients having UGIB with aspirin use had a longer length of hospital stay, but the study included those with variceal bleeding. Our study has a number of strengths. First, data collection was performed using the ICD-9/10 codes that resulted in the identification of all the potential gastrointestinal bleeding cases, after which each case was reviewed individually using well-developed prespecified criteria. Furthermore, a comprehensive structured patient form for data collection was used. Second, we had a relatively large sample size with a homogenous cause of PUD-related UGIB. Third, we included a group of patients who were not taking any antiplatelet agents or anticoagulants. Moreover, this was a comprehensive examination of the association between aspirin or anticoagulant intake and a multitude of adverse outcomes in these patients. We recorded the etiologies of death in our patients in order to understand better the association between aspirin or anticoagulant use and patients’ death. Finally, the study was conducted in a tertiary care referral center where all the diagnostic and therapeutic procedures are standardized. There are some limitations in our study. Given that the study had a retrospective design and the fact that anticoagulant intake prior to the symptom presentation depended mainly on the patients’ recall, there might be a potential recall bias. However, we implemented a rigorous history-taking system at our institution where data on medication intake was gathered separately by the admitting physician as well as by the admitting nurse, and then both data sheets were compared for discrepancies. Moreover, these data were gathered from patients and their family members and compared to their most recent hospital or clinic visits, as most of our patients are usually regular attendants of our institutions. We also recog-
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
Journal of Digestive Diseases 2014; 15; 283–292 nized that observed differences in clinical outcomes might be due to a history of antithrombotic use or their discontinuation upon admission. Another limitation was the lack of complete data on Helicobacter pylori infection and endoscopic stigmata of bleeding, as these parameters were not recorded at the time of data collection. Therefore, we are unable to report the results at admission or during long-term followup. Because our study included patients over a long period of time, changes in clinical practice might have influenced the outcomes. However, to assess this temporal variation we divided the whole cohort into three consecutive 6-year time periods and repeated the same analysis; the results were similar in each time frame (data not shown). Finally, this study was performed at a single institution; therefore, its applicability to other populations requires further evaluation. In conclusion, our study suggested that aspirin intake might be associated with favorable clinical outcomes in PUD-related UGIB, while anticoagulants seem to be associated with worse outcomes. Further investigations are needed to validate these findings. Furthermore, if aspirin is found to improve outcomes in such patients, clinicians should be informed whether and when to resume aspirin in patients taking it for primary prophylaxis.
Antithrombotics on UGIB outcomes
6
7
8
9 10 11
12
13
14
ACKNOWLEDGEMENT Part of this research was presented as a poster presentation at the Digestive Disease Week 2012 held by the American Gastroenterological Association in San Diego, California, USA. REFERENCES 1 van Leerdam ME, Vreeburg EM, Rauws EA et al. Acute upper GI bleeding: did anything change? Time trend analysis of incidence and outcome of acute upper GI bleeding between 1993/1994 and 2000. Am J Gastroenterol 2003; 98: 1494–9. 2 Wang YR, Richter JE, Dempsey DT. Trends and outcomes of hospitalizations for peptic ulcer disease in the United States, 1993 to 2006. Ann Surg 2010; 251: 51–8. 3 Marmo R, Koch M, Cipolletta L et al. Predictive factors of mortality from nonvariceal upper gastrointestinal hemorrhage: a multicenter study. Am J Gastroenterol 2008; 103: 1639–47; quiz 1648. 4 Hearnshaw SA, Logan RF, Lowe D, Travis SP, Murphy MF, Palmer KR. Acute upper gastrointestinal bleeding in the UK: Patient characteristics, diagnoses and outcomes in the 2007 UK audit. Gut 2011; 60: 1327–35. 5 Sung JJ, Tsoi KK, Ma TK, Yung MY, Lau JY, Chiu PW. Causes of mortality in patients with peptic ulcer bleeding: a
15
16 17
18
19
20
291
prospective cohort study of 10,428 cases. Am J Gastroenterol 2010; 105: 84–9. Cooper GS, Yuan Z, Rosenthal GE, Chak A, Rimm AA. Lack of gender and racial differences in surgery and mortality in hospitalized Medicare beneficiaries with bleeding peptic ulcer. J Gen Intern Med 1997; 12: 485–90. Theocharis GJ, Thomopoulos KC, Sakellaropoulos G, Katsakoulis E, Nikolopoulou V. Changing trends in the epidemiology and clinical outcome of acute upper gastrointestinal bleeding in a defined geographical area in Greece. J Clin Gastroenterol 2008; 42: 128–33. García Rodríguez LA, Lin KJ, Hernández-Díaz S, Johansson S. Risk of upper gastrointestinal bleeding with low-dose acetylsalicylic acid alone and in combination with clopidogrel and other medications. Circulation 2011; 123: 1108–15. Sostres C, Lanas A. Should prophylactic low-dose aspirin therapy be continued in peptic ulcer bleeding? Drugs 2011; 71: 1–10. Sung JJ, Lau JY, Ching JY et al. Continuation of low-dose aspirin therapy in peptic ulcer bleeding: a randomized trial. Ann Intern Med 2010; 152: 1–9. Lanas A, Aabakken L, Fonseca J et al. Clinical predictors of poor outcomes among patients with nonvariceal upper gastrointestinal bleeding in Europe. Aliment Pharmacol Ther 2011; 33: 1225–33. Åhsberg K, Höglund P, Staël von Holstein C. Mortality from peptic ulcer bleeding: the impact of comorbidity and the use of drugs that promote bleeding. Aliment Pharmacol Ther 2010; 32: 801–10. Marmo R, Koch M, Cipolletta L et al.; Italian registry on upper gastrointestinal bleeding (Progetto Nazionale Emorragie Digestive–PNED 2). Predicting mortality in non-variceal upper gastrointestinal bleeders: validation of the Italian PNED score and prospective comparison with the Rockall score. Am J Gastroenterol 2010; 105: 1284–91. Mose H, Larsen M, Riis A, Johnsen SP, Thomsen RW, Sørensen HT. Thirty-day mortality after peptic ulcer bleeding in hospitalized patients receiving low-dose aspirin at time of admission. Am J Geriatr Pharmacother 2006; 4: 244–50. Ortiz V, Ortuño J, Rodríguez-Soler M, Iborra M, Garrigues V, Ponce J. Outcome of non-variceal acute upper gastrointestinal bleeding in patients with antithrombotic therapy. Digestion 2009; 80: 89–94. Choudari CP, Rajgopal C, Palmer KR. Acute gastrointestinal haemorrhage in anticoagulated patients: diagnoses and response to endoscopic treatment. Gut 1994; 35: 464–6. Thomopoulos KC, Mimidis KP, Theocharis GJ, Gatopoulou AG, Kartalis GN, Nikolopoulou VN. Acute upper gastrointestinal bleeding in patients on long-term oral anticoagulation therapy: endoscopic findings, clinical management and outcome. World J Gastroenterol 2005; 11: 1365–8. Wolf AT, Wasan SK, Saltzman JR. Impact of anticoagulation on rebleeding following endoscopic therapy for nonvariceal upper gastrointestinal hemorrhage. Am J Gastroenterol 2007; 102: 290–6. Barada K, Abdul-Baki H, El Hajj II, Hashash JG, Green PH. Gastrointestinal bleeding in the setting of anticoagulation and antiplatelet therapy. J Clin Gastroenterol 2009; 43: 5–12. Hallas J, Dall M, Andries A et al. Use of single and combined antithrombotic therapy and risk of serious upper gastrointestinal bleeding: population based case-control study. BMJ 2006; 333: 726.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
292
H Abu Daya et al.
21 Pirmohamed M, James S, Meakin S et al. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ 2004; 329: 15–9. 22 Taha AS, Angerson WJ, Knill-Jones RP, Blatchford O. Clinical outcome in upper gastrointestinal bleeding complicating low-dose aspirin and antithrombotic drugs. Aliment Pharmacol Ther 2006; 24: 633–6. 23 Wong GL, Wong VW, Chan Y et al. High incidence of mortality and recurrent bleeding in patients with Helicobacter pylori-negative idiopathic bleeding ulcers. Gastroenterology 2009; 137: 525–31. 24 Chiu PW, Ng EK, Cheung FK et al. Predicting mortality in patients with bleeding peptic ulcers after therapeutic
Journal of Digestive Diseases 2014; 15; 283–292 endoscopy. Clin Gastroenterol Hepatol 2009; 7: 311–6; quiz 253. 25 Marmo R, Del Piano M, Rotondano G et al. Mortality from nonulcer bleeding is similar to that of ulcer bleeding in high-risk patients with nonvariceal hemorrhage: a prospective database study in Italy. Gastrointest Endosc 2012; 75: 263–72. 26 Rubin TA, Murdoch M, Nelson DB. Acute GI bleeding in the setting of supratherapeutic international normalized ratio in patients taking warfarin: endoscopic diagnosis, clinical management, and outcomes. Gastrointest Endosc 2003; 58: 369–73.
© 2014 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and Wiley Publishing Asia Pty Ltd
