The risk factors for gastrointestinal bleeding (GIB) in clopidogrel users have not been identified.
The risk factors for gastrointestinal bleeding (GIB) in clopidogrel users have not been identified.
To clarify whether clopidogrel use is a risk factor for upper GIB (UGIB) and lower GIB (LGIB) and identify the risk factors in clopidogrel users.
Using the National Health Insurance Research Database of Taiwan, 3238 clopidogrel users and 12 952 age-, sex-, and enrolment time-matched controls in a 1:4 ratio were extracted for comparison from a cohort dataset of 1 000 000 randomly sampled subjects. Cox proportional hazard regression models were used to identify the independent risk factors for UGIB and LGIB in all enrollees and clopidogrel users after adjustments for age, gender, comorbidity [i.e., coronary artery disease, hypertension, diabetes, chronic obstructive pulmonary disease, chronic kidney disease (CKD), cirrhosis, uncomplicated peptic ulcer disease, and peptic ulcer bleeding (PUB)], and medications [e.g., nonsteroidal anti-inflammatory drugs (NSAIDs), cyclooxygenase-2 inhibitors, aspirin, steroids, selective serotonin reuptake inhibitors (SSRIs), warfarin and alendronate].
Cox proportional hazard regression analysis showed that use of clopidogrel increased the risk of UGIB [hazard ratio (HR): 3.66; 95% confidence interval (CI): 2.96–4.51] and LGIB [HR: 3.52, 95% CI: 2.74–4.52]. Age, CKD, PUB history, use of aspirin and NSAIDs were independent risk factors for UGIB in the clopidogrel users. Age, CKD, PUB history, use of aspirin and SSRIs were independent risk factors for LGIB.
In clopidogrel users, age, CKD, PUB history, use of aspirin and NSAIDs are independent risk factors for UGIB; age, CKD, PUB history, use of aspirin and SSRIs are independent risk factors for LGIB.
The major concern in the use of aspirin for the prevention of cardiovascular (CV) events is the development of peptic ulcer disease (PUD), including peptic ulcer bleeding (PUB). The patients who have a past history of PUD or its complications, who take larger doses of aspirin or combine clopidogrel with aspirin, who receive concomitant steroids, nonsteroidal anti-inflammatory drugs (NSAIDs) or anti-coagulants, and who have Helicobacter pylori infection are all at higher risk of developing aspirin-associated PUB.[2, 3] Clopidogrel is an alternative anti-platelet agent that inhibits adenosine diphosphate-induced platelet aggregation without inhibiting cyclooxygenase function and prostaglandin formation. It causes less upper gastrointestinal bleeding (UGIB) and has a higher CV safety and is commonly used for secondary CV event prevention in place of aspirin in patients who have experienced aspirin-related GI adverse events or who have an allergy to aspirin. Although clopidogrel is not safe enough for patients with aspirin-related PUB due to recurrent UGIB,[5, 6] anti-secretory agents such as proton pump inhibitors (PPI) can effectively decrease aspirin or clopidogrel-related UGIB.[5, 7-9]
Few studies evaluating the risk factors for UGIB and lower GIB (LGIB) in clopidogrel users have considered confounding factors such as underlying comorbidities and other ulcerogenic medications. This nationwide cohort study aimed to identify whether the use of clopidogrel increased the risk of UGIB and LGIB compared with subjects without anti-platelets, and to identify the risk factors for UGIB and LGIB in clopidogrel users after adjusting for age, gender, underlying comorbidities and certain medications.
The National Health Insurance (NHI) programme in Taiwan was initiated in 1995, and currently covers over 99% of the 23 million population of Taiwan. The NHI research database (NHIRD) is one of largest administrative healthcare databases in the world. The present study analysed data from the Longitudinal Health Insurance Database (LHID2000) of the National Health Research Institute, which includes a cohort dataset of 1 000 000 randomly sampled individuals who were still alive in 2000. All enrollees were traced retrospectively to 1996 and followed up to 2010. There are no statistically significant differences in age, sex and healthcare cost distribution between the patients in the LHID2000 and the original NHIRD. The comprehensive healthcare data contained enrolment files, claims data, catastrophic illness files and registry for drug prescription (i.e. dose, frequency, administration route, start and end dates).
In the cohort dataset, the primary identification number of each patient was encrypted for privacy. Therefore, as this cohort dataset consisted of de-identified secondary data released to the public for research purposes, informed consent from the patients was not required. The study was approved by the Institutional Review Board of Taipei Veterans General Hospital.
According to the guidelines of clopidogrel usage issued by the Taiwan NHI, the cost of clopidogrel is covered for the patients who have a history of PUD within the past year, who are intolerant or allergic to aspirin, or who have taken aspirin concomitantly after acute coronary syndrome or after coronary interventions with stenting. Patients who took clopidogrel with an average dose of more than 150 defined daily dose per one-half year were considered for enrolment into the clopidogrel group after excluding those who took ticlopidine during the study period. Use of aspirin was allowed. After excluding the patients with alcohol-related diseases [International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes: 291.xx, 303.xx, 305.xx, 571.0, 571.1, 571.2, 571.3], malignancy of the GI tract (150.xx, 151.xx, 152.xx, 153.xx, 154.xx), inflammatory bowel disease (556.x, 555.x), coagulopathy (286.xx), vascular insufficiency of the intestine (557.xx) and gastroenteritis or colitis due to radiation (558.1) before January 1, 2000, 3238 clopidogrel users were identified from the LHID2000. The numbers of enrolled clopidogrel users in each year were 41, 232, 469, 396, 278, 329, 357, 426, 416 and 294 from 2001 to 2010. Censoring occurred when the study subjects (i) withdrew from the NHI programme (including those who died) or (ii) took less than an average clopidogrel dose of 150 DDD per one-half year. The median follow-up duration in the clopidogrel users was 1.30 years (mean: 2.07 years; range 0.01–9.81 years).
Using the same exclusion criteria as those for the study group, a control group of 12 952 subjects who had not taken any anti-platelets such as aspirin, ticlopidine or clopidogrel before and after enrolment were selected from the same cohort dataset at a 1:4 ratio (study:controls). They were matched with the study group in terms of age, gender and time of enrolment.
Other recorded covariates included age, sex, pre-existing hypertension (ICD-9-CM codes 401.xx–405.xx), diabetes mellitus (ICD-9-CM codes 250.xx), coronary artery disease (ICD-9-CM codes 411.xx–414.xx), chronic obstructive pulmonary disease (ICD-9-CM codes 491.xx, 492.xx, 494.xx, and 496.xx), chronic kidney disease (CKD) (ICD-9-CM codes 585, 586, 588.8,588.9, 250.4, 274.1, 403.x1, 404.x2, 404.x3, and 440.1), liver cirrhosis (ICD-9-CM codes 571.2, 571.5, and 571.6), ischaemic stroke (ICD-9-CM codes 433.xx, 434.xx), dyslipidaemia (ICD-9-CM codes 272.0, 272.01, 272.3, 272.4), uncomplicated PUD (ICD-9-CM codes 531.30, 531.70, 531.90, 532.30, 532.70, 532.90, 533.30, 533.70, and 533.90) and PUB (ICD-9-CM codes 531.0, 531.00, 531.01, 531.2x, 531.4x, 531.6x, 532.0, 532.00, 532.01, 532.2x, 532.4x, 532.6x, 533.00, 533.01, 533.2x, 533.4x, 533.6x, 534.0, 534.00, 534.01, 534.2x, 534.4x and 534.6x). To be designated as having a certain comorbidity, the subjects must have been given a corresponding ICD-9-CM code as their primary hospitalisation claim once or three times for out-patient claims before enrolment.
Medication [aspirin (low-dose), NSAIDs (oral or parenteral), selective cyclooxygenase-2 inhibitors (COX-2 inhibitors), steroids (oral or parenteral), clopidogrel, ticlopidine, warfarin, alendronate, selective serotonin reuptake inhibitors (SSRIs) and histamine 2 receptor antagonists (H2RAs)] were identified and classified according to the National Drug Code and the Anatomic Therapeutic Chemical codes, which are internationally accepted classification systems of drugs coordinated by the World Health Organization Collaborating Center for Drug Statistics Methodology.[10-12] Use of the aforementioned medications was defined as a prescription of these medication for more than 4 weeks within 8 weeks before the index date.
The primary end point of this study was the occurrence of nonvariceal UGIB as the main diagnosis (the top three diagnoses) during hospitalisation after January 1, 2000 (ICD-9-CM codes 530.7, 530.82, 531.0, 531.00, 531.01, 531.2, 531.2x, 531.4, 531.4x, 531.6, 531.6x, 532.0, 532.00, 532.01, 532.2, 532.2x, 532.4, 532.4x, 532.6, 532.6x, 533.0, 533.00, 533.01, 533.2, 533.2x, 533.4, 533.4x, 533.6, 533.6x, 534.0, 534.00, 534.01, 534.2, 534.2x, 534.4, 534.4x, 534.6, 534.6x, 535.X1, 537.83, 537.84, 578.0). The secondary end point was the occurrence of LGIB (ICD-9-CM 562.02, 562.03, 562.12, 562.13, 569.86, 569.3, 569.85, 578.1, 578.9) as the main diagnosis (the top three diagnoses) during hospitalisation after January 1, 2000. The two end points were followed and analysed respectively and independently.
All data were expressed as frequency (percentage) or mean ± s.d. Parametric continuous data between the study and control groups were compared by the Student's t-test, while categorical data were compared by the χ2 test or Fisher's exact test, as appropriate, by the Spss software. Yates' correction was not used. The cumulative hazard was assessed using Kaplan–Meier analysis, with significance based on the log-rank test.
Multiple regression analysis was conducted using Cox proportional hazard regression analysis to identify the risk factors for UGIB and LGIB. Statistical significance was considered at a two-sided P value of less than 0.05. Microsoft SQL Server 2005 was used for data management and computing. All statistical analyses were performed using the Spss software package (15.0, SPSS Inc., Chicago, IL, USA).
The demographic data of the control group and clopidogrel group are shown in Table 1. There were no significant differences between the groups in age, gender, rate of cirrhosis and the use of NSAIDs and alendronate. However, the clopidogrel group had significantly higher rates of hypertension, coronary heart disease, diabetes, chronic obstructive pulmonary disease, CKD, ischaemic stroke, uncomplicated PUD, PUB, and dyslipidaemia, and use of COX-2 inhibitors, aspirin, steroids, SSRIs, warfarin and H2RAs (Table 1).
|Controls N = 12 952||Clopidogrel N = 3238||P value|
|Age, years||68.56 ± 11.62||68.56 ± 11.62||0.992|
|Male, n (%)||7596 (58.65)||1899 (58.65)||1.000|
|Coronary artery disease, n (%)||1638 (12.65)||2145 (66.24)||<0.001|
|Hypertension, n (%)||6456 (49.85)||2760 (85.24)||<0.001|
|Diabetes, n (%)||2797 (21.60)||1447 (44.69)||<0.001|
|Chronic obstructive pulmonary disease, n (%)||3395 (26.21)||1071 (33.08)||<0.001|
|Chronic renal disease, n (%)||912 (7.04)||611 (18.87)||<0.001|
|Cirrhosis, n (%)||329 (2.54)||73 (2.25)||0.192|
|Ischaemic stroke, n (%)||460 (3.55)||1109 (34.25)||<0.001|
|Uncomplicated peptic ulcer disease, n (%)||4285 (33.08)||1638 (50.59)||<0.001|
|Peptic ulcer bleeding, n (%)||943 (7.28)||451 (13.93)||<0.001|
|Dyslipidaemia, n (%)||3496 (26.99)||1706 (52.69)||<0.001|
|NSAIDs, n (%)||2484 (19.18)||647 (19.98)||0.125|
|Cyclooxygenase-2 inhibitors, n (%)||697 (5.38)||236 (7.29)||<0.001|
|Aspirin, n (%)||0 (0.00)||428 (13.22)|
|Steroids, n (%)||629 (4.86)||189 (5.84)||0.012|
|SSRIs, n (%)||125 (0.97)||52 (1.61)||0.001|
|Warfarin, n (%)||42 (0.32)||51 (1.58)||<0.001|
|Alendronate, n (%)||85 (0.66)||21 (0.65)||0.336|
|H2RAs, n (%)||874 (6.75)||277 (8.55)||<0.001|
|Upper gastrointestinal bleeding, n (%)||357 (2.76)||255 (7.88)||<0.001|
|Lower gastrointestinal bleeding, n (%)||263 (2.03)||170 (5.25)||<0.001|
During the 10-year follow-up period, 612 (3.78%) of the 16 190 subjects developed UGIB and 433 (2.67%) developed LGIB (Table 1). The cumulative hazard of UGIB and LGIB using Kaplan–Meier analysis showed that the clopidogrel group had a significantly higher hazard for UGIB and LGIB than the control group (all P < 0.001) (Figure 1).
After adjusting for age, gender, presence of coronary artery disease, hypertension, diabetes, chronic obstructive pulmonary disease, CKD, cirrhosis, ischaemic stroke, uncomplicated PUD, PUB, dyslipidaemia, and the use of NSAIDs, COX-2 inhibitors, aspirin, clopidogrel, steroids, SSRIs, warfarin, alendronate and H2RAs, use of clopidogrel was found to be an independent risk factor for UGIB [hazard ratio (HR): 3.66, 95% confidence interval (CI): 2.96–4.51] and LGIB (HR: 3.52, 95% CI: 2.74–4.52). The other independent risk factors for UGIB were age, male gender, CKD, cirrhosis, uncomplicated PUD, PUB, use of NSAIDs, COX-2 inhibitors, aspirin, clopidogrel, steroids and H2RAs. The other independent risk factors for LGIB were age, hypertension, diabetes, CKD, cirrhosis, ischaemic stroke, PUB and use of aspirin.
The role of age and concomitant use of aspirin in the clopidogrel users was then evaluated. Figures 2 and 3 show the cumulative hazard of UGIB and LGIB in the patients who were ≥75 years old and <75 years old, respectively, and were grouped by clopidogrel with aspirin, clopidogrel without aspirin and matched controls. The patients taking clopidogrel with aspirin had a significantly higher risk of UGIB and LGIB than the patients taking clopidogrel without aspirin (P < 0.05 by the log-rank test). The patients taking clopidogrel without aspirin had a significantly higher risk of UGIB and LGIB than the matched controls (P < 0.05).
After adjusting for age, gender, the aforementioned comorbidities and medications, age, CKD, PUB history and concomitant use of aspirin or NSAIDs were independent risk factors for UGIB in the clopidogrel users (Table 2). Age, CKD, history of PUB and concomitant use of aspirin or SSRIs were independent risk factors for LGIB in the clopidogrel users (Table 3). Further analysis showed that age (HR: 1.04, 95% CI: 1.01–1.08), uncomplicated PUD (HR: 2.07, 95% CI: 1.04–4.10), PUB (HR: 4.04, 95% CI: 2.06–7.92) and the use of NSAIDs (HR: 1.81, 95% CI: 1.02–3.66) were independent risk factors for UGIB in the patients taking clopidogrel with aspirin. In addition, CKD (HR: 1.65, 95% CI: 1.02–3.38) was an independent risk factor for LGIB in the patients taking clopidogrel with aspirin.
|Univariate HR (95% CI)||P value||Multivariate HRa (95% CI)||P value|
|Age||1.03 (1.02–1.04)||<0.001||1.03 (1.02–1.05)||<0.001|
|Male||1.00 (0.79–1.29)||0.952||1.08 (0.83–1.39)||0.573|
|Coronary artery disease||1.57 (1.19–2.08)||<0.001||1.28 (0.95–1.71)||0.101|
|Hypertension||1.96 (1.26–3.04)||0.001||1.23 (0.78–1.93)||0.373|
|Diabetes||1.21 (0.95–1.55)||0.132||1.04 (0.80–1.35)||0.782|
|COPD||1.26 (0.98–1.63)||0.078||0.89 (0.68–1.17)||0.399|
|Chronic renal disease||1.83 (1.38–2.43)||<0.001||1.58 (1.17–2.13)||0.003|
|Cirrhosis||2.56 (1.43–4.58)||0.001||1.74 (0.96–3.17)||0.070|
|Ischaemic stroke||1.07 (0.83–1.38)||0.602||1.10 (0.84–1.43)||0.464|
|Uncomplicated peptic ulcer disease||1.67 (1.29–2.16)||<0.001||1.13 (0.86–1.50)||0.383|
|Peptic ulcer bleeding||4.13 (3.21–5.33)||<0.001||3.70 (2.82–4.85)||<0.001|
|Dyslipidaemia||0.74 (0.58–0.95)||0.017||0.65 (0.50–0.84)||0.001|
|NSAIDs||1.37 (1.10–1.70)||0.025||1.19 (1.09–1.33)||0.019|
|Cyclooxygenase-2 inhibitor||1.54 (1.05–2.27)||0.027||1.26 (0.85–1.88)||0.253|
|Aspirin||2.88 (2.11–3.93)||<0.001||2.98 (2.17–4.11)||<0.001|
|Steroid||1.18 (0.66–2.12)||0.576||1.30 (0.84–2.02)||0.237|
|SSRI||2.33 (1.33–5.26)||0.040||1.27 (0.56–2.89)||0.565|
|Warfarin||1.27 (0.43–3.04)||0.813||1.14 (0.50–2.62)||0.752|
|Alendronate||1.40 (0.45–4.38)||0.568||0.99 (0.31–3.17)||0.990|
|H2RA||0.88 (0.49–1.58)||0.659||1.06 (0.71–1.56)||0.780|
|Univariate HR (95% CI)||P value||Multivariate HRa (95% CI)||P value|
|Age||1.06 (1.04–1.08)||<0.001||1.05 (1.04–1.07)||<0.001|
|Male||0.91 (0.67–1.22)||0.528||0.99 (0.72–1.35)||0.930|
|Coronary artery disease||1.17 (0.85–1.61)||0.340||1.03 (0.73–1.45)||0.864|
|Hypertension||1.58 (0.97–2.58)||0.066||1.26 (0.76–2.10)||0.371|
|Diabetes||1.19 (0.88–1.62)||0.344||1.18 (0.85–1.65)||0.313|
|COPD||1.70 (1.26–2.31)||0.001||1.25 (0.90–1.72)||0.180|
|Chronic renal disease||1.85 (1.31–2.62)||0.001||1.70 (1.18–2.46)||0.004|
|Cirrhosis||2.05 (0.90–4.63)||0.086||1.92 (0.84–4.39)||0.125|
|Ischaemic stroke||1.24 (0.92–1.69)||0.161||1.25 (0.91–1.71)||0.171|
|Uncomplicated peptic ulcer disease||1.20 (0.88–1.62)||0.246||0.91 (0.66–1.26)||0.570|
|Peptic ulcer bleeding||1.68 (1.16–2.44)||0.006||1.60 (1.08–2.36)||0.019|
|Dyslipidaemia||0.76 (0.56–1.03)||0.073||0.79 (0.57–1.09)||0.163|
|NSAIDs||0.81 (0.66–1.12)||0.099||0.76 (0.45–1.19)||0.111|
|Cyclooxygenase-2 inhibitor||0.53 (0.25–1.21)||0.100||0.47 (0.22-1.02)||0.054|
|Aspirin||2.53 (1.72–3.76)||<0.001||2.83 (1.89–4.24)||<0.001|
|Steroid||1.18 (0.66–2.12)||0.576||1.23 (0.68–2.25)||0.493|
|SSRI||2.33 (1.33–5.26)||0.040||2.22 (1.01–4.58)||0.048|
|Warfarin||1.27 (0.42–3.04)||0.813||0.97 (0.35–2.67)||0.953|
|Alendronate||1.40 (0.45–4.38)||0.568||1.52 (0.47–4.89)||0.484|
|H2RA||0.88 (0.49–1.58)||0.659||0.72 (0.39–1.31)||0.282|
In this study, the long-term use of clopidogrel incurred a significantly higher risk of developing UGIB and LGIB in the general population after adjusting for age, gender, comorbidity (i.e., coronary artery disease, hypertension, diabetes, chronic obstructive pulmonary disease, chronic renal disease, cirrhosis, uncomplicated PUD, PUB and dyslipidaemia) and medications (e.g., NSAIDs, COX-2 inhibitors, aspirin, steroids, clopidogrel, SSRIs, warfarin, and alendronate). Age, CKD, PUB history, use of aspirin and NSAIDs are independent risk factors for UGIB in the clopidogrel users. Age, CKD, PUB history and use of aspirin and SSRIs are independent risk factors for LGIB in the clopidogrel users.
According to the guidelines of clopidogrel usage issued by the Taiwan NHI, most prescriptions of clopidogrel are for alternative anti-platelet therapy in place of aspirin in patients who have experienced PUD or PUB and dual anti-platelet therapy for acute coronary syndrome or after coronary interventions. The results showed that PUB and concomitant use of aspirin, but not uncomplicated PUD, were important risk factors for UGIB in clopidogrel users, which is consistent with previous studies showing that concomitant use of aspirin or a history of PUB increases the risk of PUB, and that clopidogrel usage is not safe due to UGIB in high-risk patients with previous PUB.[5, 13, 14] The finding that clopidogrel may delay gastric ulcer healing provides a reasonable explanation why PUD or PUB reoccurs in clopidogrel-naïve users after previous PUD has healed.[5, 6, 15] Although other studies have reported that the use of clopidogrel did not increase risk of UGIB and LGIB in patients with CKD or renal failure,[10, 12] the current study showed that CKD including chronic renal failure (ICD-9-CM codes 585) was an important risk factor for both UGIB and LGIB in clopidogrel users.
Helicobacter pylori infection is an important risk factor for PUB and UGIB. The NHIRD does not contain data on the real status of H. pylori infection, although most PUD/PUB patients had H. pylori eradicated, which is covered by the Taiwan NHI. However, previous studies have shown that H. pylori infection or H. pylori eradication does not have an important impact on PUB in clopidogrel users.[5, 6, 13] Another issue is ulcer prophylaxis. Previous studies have shown that PPIs rather than H2RAs can prevent clopidogrel-related PUD/PUB.[17, 18] However, the use of PPIs is strictly limited by the NHI for patients with PUD/PUB or endoscopic erosive esophagitis for 4 months or more.[11, 19] Gastroprotective agents for prophylaxis against ulcers or UGIB are not paid for by the NHI, and prophylactic prescriptions of PPIs including over-the-counter medications are therefore not found in the NHIRD. Only generic (cheaper) H2RAs are covered for prophylaxis against ulcers or UGIB. Our data suggest that although H2RAs were prescribed to high-risk patients for UGIB prophylaxis, they did not seem to prevent the occurrence of UGIB.
In general, LGIB is defined as GI bleeding distal to the ligament of Treitz. The definition of LGIB according to ICD-9-CM coding as used in the current study included diverticulosis/diverticulitis of the small intestine with haemorrhage, diverticulosis/diverticulitis of the colon with haemorrhage, angiodysplasia of the intestine with haemorrhage, Dieulafoy lesions (haemorrhage) of the intestine, haemorrhage of the rectum and anus, blood in stools, and unspecified haemorrhage of the GI tract. A few previous studies have evaluated the risk factors for LGIB in the general population or in subgroup analysis; however, no study has evaluated the risk factors for LGIB in clopidogrel users. The current study found that the use of clopidogrel was an independent risk factor for LGIB, and old age, CKD, PUB history and concomitant use of aspirin or SSRIs were independent risk factors for LGIB in clopidogrel users. This is consistent with a previous study in which a combination of aspirin, clopidogrel and SSRI use increased the risk of bleeding in post-myocardial infarction patients. The possible mechanisms by which CKD increases both UGIB and LGIB in clopidogrel users include platelet dysfunction, abnormalities in blood coagulation and platelet-vessel wall interactions in patients with chronic renal disease. In addition, age and CKD are important risk factors for angiodysplasia bleeding, which can occur in UGI and LGI. The use of aspirin also increased both UGIB and LGIB.[7, 22]
There are several limitations to this study. First, clopidogrel is not a first-line anti-platelet agent in Taiwan, and is only covered by the Taiwan NHI for the patients who have a history of PUD/PUB, who are allergic to aspirin, and who are under dual therapy with aspirin after acute coronary syndrome or after coronary interventions. About 65% of the clopidogrel users in the current study had a history of uncomplicated PUD or PUB, and about 13% of the clopidogrel users took aspirin concomitantly. Certain selection bias and channelling bias exist in this study such that caution must be taken in extrapolating the results. In contrast to the clopidogrel users who had some risks for GIB, the control group did not take any anti-platelets such as aspirin, so the HRs of UGI and LGIB in the clopidogrel group were overestimated. However, the key point of this study is to identify the possible risk factors of UGIB and LGIB in the clopidogrel users and not in all enrollees. Second, although H. pylori infection, smoking and drinking are also important risk factors for UGIB, such information is not available from the NHIRD. Third, we assumed that the use of over-the-counter medications (such as aspirin, clopidogrel, NSAIDs, and PPIs), details of which were unavailable in the dataset, was limited, because more than 99% of the residents in Taiwan are covered by the NHI, which is easily accessible and offers low copayments to the general population.[10, 19, 23] However, over-the-counter medications had an important impact in our data interpretation. This could not be fully avoided and could not be validated in this epidemiological study.
The use of clopidogrel did increase risks of both UGIB and LGIB after adjusting for age, gender, comorbidity and some medications. Age, CKD, PUB history and use of aspirin and NSAIDs are important risk factors for UGIB in clopidogrel users. Age, CKD, PUB history and use of aspirin and SSRI are important risk factors for LGIB in clopidogrel users. H2RA therapy did not protect against UGIB in the clopidogrel users. However, the benefits of prophylactic PPIs in the clopidogrel users could not be confirmed in the NHRI database as the NHI system does not pay for the use of prophylactic PPIs.
Guarantor of the article: Jiing-Chyuan Luo, M.D.
Author contributions: Chung-Chi Lin: collected, analysed and interpreted the data; drafted and revised the manuscript. Jiing-Chyuan Luo: planned and conducted the study; interpreted the data; drafted and revised the manuscript; has full responsibility for the conduct of the study and control of the decision to publish the study. Hsiao-Yun Hu and Yen-Ling Peng: collected, analysed, and interpreted the data; revised the manuscript. Ming-Chih Hou, Han-Chieh Lin, Fa-Yauh Lee: analysed/interpreted the data; revised the manuscript. All authors approved the final version of the manuscript.
Declaration of personal interests: None.
Declaration of funding interests: This study was funded in part by Taipei Veterans General Hospital (V101C-028, V102C-006) and in part by National Science Council of Taiwan (NSC 101-2314-B-010-012 -MY3). The initial data analyses or writing of this paper was not funded by other grants.