SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

Background

Peptic ulcer bleeding remains a major healthcare problem despite decreasing prevalence of peptic ulcer disease. The role of chronic obstructive pulmonary disease (COPD) in the risk of peptic ulcer bleeding has not yet been established.

Aim

To determine if COPD patients have a higher risk of peptic ulcer bleeding than the general population and to identify the risk factors of peptic ulcer bleeding in COPD patients.

Methods

From Taiwan's National Health Insurance research database, 62 876 patients, including 32 682 COPD and 30 194 age-gender-matched non-COPD controls, were recruited. Cox proportional hazard regression was performed to evaluate independent risk factors for ulcer bleeding in all patients and to identify risk factors in COPD patients.

Results

During the 8-year follow-up, COPD patients had a significant higher rate of peptic ulcer bleeding than the control group (< 0.001, by log-rank test). By Cox proportional hazard regression analysis, COPD [hazard ratio (HR) 1.93, 95% CI 1.73–2.17] was an independent risk factor after adjusting for age, gender, underlying comorbidities and ulcerogenic medication. Age > 65 years, male, comorbidities of hypertension, diabetes, heart failure, history of peptic ulcer disease, and chronic renal disease and use of nonsteroidal anti-inflammatory drugs were risk factors of ulcer bleeding in COPD patients.

Conclusion

Patients with chronic obstructive pulmonary disease have a higher risk of peptic ulcer bleeding after adjustments for possible confounding factors like underlying comorbidities and ulcerogenic medication.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

Although Helicobacter pylori (H. pylori) infection and the use of nonsteroidal anti-inflammatory drugs (NSAIDs) play important roles in peptic ulcer formation and bleeding,[1] peptic ulcer disease is not uncommon in patients with chronic medical conditions like liver cirrhosis, renal failure and chronic obstructive pulmonary disease (COPD). In literature, patients who smoke or have COPD carry higher risk of developing peptic ulcer disease.[2-4] Moreover, COPD patients often have other smoking-related chronic diseases, such as hypertension, coronary artery disease, or heart failure, and use similar medications, including anti-platelet drugs or corticosteroids. These drugs may be ulcerogenic or delay ulcer healing.[5-7]

Peptic ulcer bleeding remains a major healthcare problem[8] with substantial economic impact[9, 10] despite decreasing peptic ulcer disease prevalence.[11, 12] To date, few population-based studies have evaluated whether COPD really increases peptic ulcer bleeding compared to the general population. Risk factors in COPD patients are also not identified. This nationwide population-based study aimed to investigate the association between COPD and peptic ulcer bleeding after adjusting for potential confounding factors such as comorbidity and ulcerogenic medication, and to identity independent risk factors for ulcer bleeding in COPD patients.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

Data sources

The National Health Insurance (NHI) programme in Taiwan, launched in 1995, covers over 99% of Taiwan's 23-million population. The NHI research database (NHIRD), established by cooperation of the Bureau of National Health Insurance (BNHI) and the National Health Research Institute (NHRI), is one of largest administrative healthcare databases in the world and is open to scientists for research purposes. The present study analysed data from the Longitudinal Health Insurance Database (LHID2000) of the NHIRD. The LHID2000 contained a cohort dataset of 1 000 000 randomly sampled individuals who were alive in 2000. All enrollees were traced retrospectively to 1996 and followed up to 2007.[13] There were no statistically significant differences in age, gender and healthcare costs distribution between the patients in the LHID2000 and the original NHIRD. Comprehensive healthcare data included enrolment files, claims data, catastrophic illness files and registry for drug prescription.

In the cohort dataset, each patient's original identification number was encrypted for privacy. Because this cohort dataset consisted of de-identified secondary data released to the public for research purposes, this study was exempted from full review by the Institutional Review Board.

Study group

The study group was composed of COPD patients (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM]) codes 491.xx, 492.xx, 494.xx and 496.xx) identified after 1 January 2000. Asthma patients (ICD-9-CM code 493.xx) and those with of any gastrointestinal tract bleeding (ICD-9-CM codes 530.82,531.xx,532.xx,533.xx, 534.xx,535.x1, 537.82, 537.83, 562.02, 562.03, 562.12, 562.13, 569.3, 569.82, 569.84, 569.85, 578.x, 456.0, 456.1, 456.20, 456.21, 530.1, 530.10, 530.11 and 530.2) as the major diagnosis during hospitalisation before 1 January 2000 were excluded.

Control group

The control group was composed of patients without COPD or asthma before and after enrolment from the database. They were matched with the study subjects in a 1:1 ratio and terms of age, gender and ulcerogenic medication (e.g. NSAIDs, acetylsalicylic acid [ASA], steroids, clopidogrel, ticlopidine and warfarin).

Other recorded covariates included age, gender, coronary artery disease (ICD-9-CM codes: 410.xx to 414.xx), hypertension (ICD-9-CM codes: 401.xx to 405.xx), diabetes mellitus (ICD-9-CM codes 250.xx), heart failure (ICD-9-CM codes 428.00) and chronic kidney disease (ICD-9-CM codes 585, 586, 588.8,588.9, 250.4, 274.1, 403.x1, 404.x2, 404.x3 and 440.1).

Incidence of hospitalisation for ulcer bleeding

The primary endpoint was the occurrence of administrative claims of ulcer bleeding as the major diagnosis during hospitalisation. Ulcer bleeding was proven by endoscopic examination, with 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, 533.0, 533.00, 533.01, 533.2x, 533.4x, 533.6x, 534.0, 534.00, 534.01, 534.2x, 534.4x and 534.6x.

Statistical analysis

Microsoft SQL Server 2005 was used for data management and computing. All statistical analyses were performed using the SPSS software (Version 18.0, SPSS Inc., Chicago, IL, USA). Demographic data were expressed as frequency (percentage) or as mean ± standard deviation. Parametric continuous data between the case and control groups were compared by Student's t-test while categorical data were compared by Chi-square test and Yates’ correction or Fisher's exact test, as appropriate. Survival analysis was assessed using the Kaplan–Meier analysis, with significance based on log-rank test. Survival time was calculated from the date of enrollment to the date of hospitalisation for ulcer bleeding. Multiple regression analysis was conducted using Cox proportional hazard regression analysis. A two-sided < 0.05 was considered statistical significant.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

Patient characteristics

The demographic data of the COPD and control groups were shown in Table 1. Of the 62 876 cases included, 32 682 were in the COPD group and 30 194 were in the control group. Concurrent use of ulcerogenic medication like ASA, NSAIDs, clopidogrel, ticlopidine and warfarin, except for steroids, was not significantly different between the two groups (> 0.05). Compared to the control group, the COPD group had higher proportion of comorbidities like hypertension, diabetes, coronary artery disease, heart failure and chronic renal disease. The COPD group also had more prevalent history of peptic ulcer disease.

Table 1. Demographic data and ulcer bleeding between patients with or without COPD
 COPDP value
No (= 30 194)Yes (= 32 682)
  1. COPD, chronic obstructive pulmonary disease; ASA, acetylsalicylic acid; NSAIDs, nonsteroidal anti-inflammatory drugs.

Age, years60.88 ± 16.6361.13 ± 16.570.053
Male, n (%)17 324 (57.4)18 809 (57.6)0.656
Hypertension, n (%)8118 (26.9)13 854 (42.4)<0.001
Diabetes, n (%)3221 (10.7)5272 (16.1)<0.001
Coronary artery disease, n (%)3192 (10.6)6921 (21.2)<0.001
Heart failure, n (%)588 (1.9)2058 (6.3)<0.001
Chronic renal disease, n (%)553 (1.8)2133 (6.5)<0.001
History of peptic ulcer disease, n (%)3068 (10.2)7207 (22.1)<0.001
Medication
ASA, n (%)4430 (14.7)4854 (14.9)0.531
NSAIDs, n (%)4526 (15.0)5009 (15.3)0.244
Steroids, n (%)460 (1.5)2013 (6.2)<0.001
Non-ASA anti-platelet agent, n (%)656 (2.2)745 (2.3)0.379
Warfarin, n (%)196 (0.6)226 (0.7)0.548
Peptic ulcer bleeding501 (1.7)944 (2.9)<0.001

Survival-free period from peptic ulcer bleeding

During the 8-year follow-up period, 1445 (2.3%) of 62 876 cases experienced ulcer bleeding, with 944 (65.3%) cases in the COPD group and 501 (34.7%) in the control group. By survival-free period from peptic ulcer bleeding events using Kaplan–Meier analysis, COPD patients had a significant higher rate of peptic ulcer bleeding than the control group (< 0.001, by log-rank test) (Figure 1)

image

Figure 1. Kaplan–Meier estimates of survival free of peptic ulcer bleeding events in subjects categorised by chronic obstructive pulmonary disease (COPD). The event-free survival rates were significantly different in two groups (< 0.001 by log-rank test).

Download figure to PowerPoint

Univariate analysis for risk factors of ulcer bleeding

Compared to non-ulcer bleeding group, patients who developed ulcer bleeding tended to be older, male, and with other comorbidities (i.e. hypertension, diabetes, coronary artery disease, heart failure, chronic renal disease and COPD), history of peptic ulcer disease, and concurrent use of ulcerogenic medication (e.g. ASA, NSAIDs, steroids, non-ASA anti-platelet agents), except for warfarin (Table 2).

Table 2. Univariate analysis for risk factors of ulcer bleeding among all the enrollees
 Peptic ulcer bleedingP value
No (= 61 431)Yes (= 1445)
  1. COPD, chronic obstructive pulmonary disease; ASA, acetylsalicylic acid; NSAIDs, nonsteroidal anti-inflammatory drugs.

Age, years60.79 ± 16.6370.42 ± 11.81<0.001
Male, n (%)35 172 (57.3)961 (66.5)<0.001
Hypertension, n (%)21 200 (34.5)772 (53.4)<0.001
Diabetes, n (%)8156 (13.3)337 (23.3)<0.001
Coronary artery disease, n (%)9721 (15.8)392 (27.1)<0.001
Heart failure, n (%)2516 (4.1)130 (9.0)<0.001
Chronic renal disease, n (%)2579 (4.2)107 (7.4)<0.001
History of peptic ulcer disease, n (%)9930 (16.2)345 (23.9)<0.001
COPD31 738 (51.7)944 (65.3)<0.001
Medication
ASA, n (%)8948 (14.6)336 (23.3)<0.001
Oral NSAIDs, n (%)9109 (14.8)426 (29.5)<0.001
Steroids, n (%)2392 (3.9)81 (5.6)0.001
Non-ASA anti-platelet agent, n (%)1331(2.2)70(4.8)<0.001
Warfarin, n (%)407 (0.7)15 (1.0)0.118

Multivariate analysis for risk factors of ulcer bleeding

After adjusting for age, gender, presence of comorbidities, history of peptic ulcer disease, and use of ulcerogenic medication, Cox regression analysis showed that old age, male gender, hypertension, diabetes, heart failure, history of peptic ulcer disease, COPD, and concurrent use of NSAIDs and non-ASA anti-platelet agents were independent risk factors of developing ulcer bleeding in all patients (Table 3).

Table 3. Independent predictors of ulcer bleeding among all the enrollees identified by Cox regression analysis
 Adjusted HR95% CIP value
  1. HR, hazard ratio; COPD, chronic obstructive pulmonary disease; ASA, acetylsalicylic acid; NSAIDs, nonsteroidal anti-inflammatory drugs.

Age1.051.04–1.05<0.001
Male1.491.34–1.67<0.001
Hypertension1.231.09–1.380.001
Diabetes1.521.33–1.72<0.001
Coronary artery disease1.050.92–1.200.496
Heart failure1.511.25–1.83<0.001
Chronic renal disease1.190.97–1.450.100
History of peptic ulcer disease1.261.10–1.43<0.001
COPD1.931.73–2.17<0.001
Medication
ASA1.060.93–1.210.362
NSAIDs1.951.74–2.19<0.001
Steroids1.060.85–1.330.610
Non-ASA anti-platelet agents1.411.11–1.800.005

Risk factors of ulcer bleeding in COPD patients

Cox multivariate regression analysis showed that age >65 years, male gender, hypertension, diabetes, heart failure, chronic renal disease, peptic ulcer history, and NSAIDs use, but not coronary artery disease and use of aspirin, steroids, non-ASA anti-platelet agents were important risk factors for ulcer bleeding in COPD patients (Table 4).

Table 4. Significant risk factors for ulcer bleeding in COPD patients, by multivariate Cox regression analysis
Risk factorsAdjusted HR95% CIP value
  1. HR, hazard ratio; COPD, chronic obstructive pulmonary disease; NSAIDs, nonsteroidal anti-inflammatory drugs.

Age >652.652.26–3.09<0.001
Male1.621.41–1.86<0.001
Hypertension1.281.10–1.480.001
Diabetes1.381.18–1.62<0.001
Heart failure1.661.34–2.06<0.001
Chronic renal disease1.301.04–1.620.019
History of peptic ulcer disease1.211.04–1.400.012
NSAIDs use1.671.44–1.93<0.001

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

In the present study, COPD patients (HR: 1.93; 95% CI, 1.73–2.17) have significantly higher risk of developing ulcer bleeding after adjustments for possible confounding factors such as age, gender, comorbidities (i.e. hypertension, diabetes, coronary heart disease, heart failure, chronic renal disease, and history of peptic ulcer disease) and ulcerogenic medication (e.g. NSAIDs, ASA, steroids, clopidogrel, ticlopidine and warfarin). Older age, male gender, hypertension, diabetes, heart failure, chronic renal disease, peptic ulcer history and NSAIDs use are important risk factors for ulcer bleeding in COPD patients.

The original study design aimed to match all major comorbid medical diseases (i.e. hypertension, diabetes, coronary heart disease, heart failure, chronic renal disease and history of peptic ulcer disease) and concurrent ulcerogenic medication between the COPD and control groups. However, it failed in comorbid disease matching because COPD patients had higher proportions of other comorbidities. This may reflect the actual real-life conditions and coincide with a recent concept that COPD is no longer a localised inflammatory disease limited to the lungs but is a systemic inflammatory disease[14, 15] that may be associated with other systemic diseases like hypertension, diabetes, cardiovascular disease or chronic kidney disease.[16-18]

During the 8-year follow-up period, COPD patients indeed had a higher risk of developing peptic ulcer bleeding than the general population (Figure 1). Previous studies also showed that smoking and COPD were risk factors of peptic ulcer disease.[2-4] The present study found that COPD patients had a higher proportion of peptic ulcer disease history (22.1 vs. 10.2%, < 0.001). Previously, most studies for evaluating the risk factors of peptic ulcer bleeding mainly focused on NSAIDs or aspirin use, or H. pylori infection. Very few studies delved on the relationship between peptic ulcer bleeding and smoking or COPD. A population-based cohort study in Denmark showed that smoking was a risk factor for ulcer perforation but not for ulcer bleeding.[19] The other case–control study found that patients with co-existing lung disease increased the risk of ulcer bleeding (OR:2.5, 95% CI 1.40–4.46) but did not specify the associated lung disease.[20] The present study is the first large-scale, longitudinal, population-based cohort study to clarify that COPD indeed increases the risk of peptic ulcer bleeding.

There may be several reasons why COPD increases ulcer bleeding risk. First, COPD is characterised not only by chronic local but also by systemic inflammation.[15] Thus, COPD patients are exposed to oxidative stress secondary to chronic hypoxia and produce reactive oxygen species (ROS)[21, 22] that may damage gastric or small intestinal mucosa.[23] Second, COPD patients share other smoking-related chronic diseases, such as hypertension, coronary artery disease, or heart failure, and have more instances using anti-platelet agents, which protect against cardiovascular events but increase ulcer bleeding risk.[24, 25] Third, COPD patients often need steroids for controlling lung inflammation. Steroid-induced peptic ulcer disease is controversial but steroids reportedly delay peptic ulcer healing.[5, 6, 26-28] Despite the multifaceted hypothesis and associated confounding factors, the present study shows that COPD (HR 1.93, 95% CI 1.73–2.17) remains an independent risk factor of ulcer bleeding after adjusting for confounding factors like hypertension, coronary artery disease, heart failure and ulcerogenic medication, including NSAIDs, anti-platelet agents and steroids.

Limitations

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

The current study has several limitations. First, this is a retrospective cohort study with observations based on hospitalised patients with peptic ulcer bleeding. Certain selection biases may exist such that caution must be taken in extrapolating the results. Second, it is well known that H. pylori infection is an important risk factor for ulcer bleeding,[1, 29] but we did not have H. pylori prevalence in the NHIR database. However, the issue whether COPD patients really have higher H. pylori seroprevalence is still controversial[30] and there even may be no association between H. pylori infection and COPD in a published study.[31] Therefore, further large-scale study is needed for evaluation the prevalence of H. pylori infection in COPD patients. Third, although smoking is an important risk factor for both COPD and peptic ulcer, such information is not available from the NHIR database. However, emerging evidence supports that risk factors other than smoking play important parts of developing COPD. An estimated 25–45% of patients with COPD have never smoked.[32] Therefore, a prospective study is warranted to clarify the role of smoking in COPD patients with ulcer bleeding. Finally, although socioeconomic status could be also a possible confounder in COPD patients with ulcer bleeding, the socioeconomic status was unavailable in our database.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

Patients with COPD have a higher risk of developing peptic ulcer bleeding than the general population. Older age, male gender, hypertension, diabetes, heart failure, chronic renal disease, peptic ulcer history and NSAIDs usage are important risk factors for ulcer bleeding in COPD patients.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References

Declaration of personal interests: None. The authors express their gratitude to Mrs. Pui-Ching Lee and Hsiu-Wen Liao (Department of Medicine, Taipei Veterans General Hospital) for her help in the statistical consultation. Declaration of funding interests: The study is funded in part by Taipei Veterans General Hospital (V99C1-015 and V100C-026).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Limitations
  8. Conclusions
  9. Acknowledgements
  10. References