SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Background

Several drug classes are known to be associated with serious upper gastrointestinal bleeding (UGIB), among others NSAID, low-dose acetylsalicylic acid (ASA), vitamin K antagonists (VKA), clopidogrel and selective serotonin reuptake inhibitors (SSRIs). There are few data on how and to what extent these drugs are reintroduced in patients who have been discharged after a bleeding episode related to any of them.

Aim

To assess if physicians re-prescribed potential causative drugs after an episode of UGIB and to explore whether drugs with antihaemostatic action (DAHA) are re-prescribed without a gastro-protective agent.

Methods

By use of the Kaplan–Meyer method, we estimated the time from UGIB to re-prescribing for 3652 cases who were all admitted to hospital with a diagnosis of serious upper gastrointestinal bleeding from 1995 to 2006. Data on drug exposure were retrieved from a Danish prescription database, a recent study on drug-related UGIB, and The National Board of Health in Denmark.

Results

One-year rates of re-prescribing after UGIB were; 82%, 25%, 43%, 68%, 55%, 71% for SSRIs, NSAID, low-dose ASA, VKA, clopidogrel and dipyridamol, respectively. However, re-prescribing rates without proton pump inhibitors (PPIs) were markedly lower 25%, 3%, 5%, 1%, 17% and 6%, respectively. Non-users of DAHA had a prevalence of PPI use of about 30% a few months after an UGIB.

Conclusions

Drugs with antihaemostatic action are re-prescribed to a large extent after an episode of upper gastrointestinal bleeding, but usually covered by PPIs. This use of PPI is specific for users of drugs with antihaemostatic action.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Upper gastrointestinal bleeding (UGIB) is a serious condition. The reported fatality rate is between 5–14%[1, 2] and recently a recurrence rate of 17.5 per 1000 person years during a mean follow-up of 3 years was reported.[3] Moreover, it is well established that gastrointestinal bleeding is the most common serious bleeding complication from the use of long-term antiplatelet therapy.[4]

Data are very scarce on how physicians re-prescribe potential causative drugs after an episode of UGIB that apparently is associated with one of these agents. Use of proton pump inhibitors (PPIs) is an effective prophylactic measure against UGIB in patients taking nonsteroidal anti-inflammatory drugs (NSAID) and ASA.[5] If potentially causative drugs are re-prescribed to a large extent, it would also be of interest whether this group of patients are prescribed PPIs.

We undertook this study to describe the extent of re-prescribing of causative agents in a large group of drug-related UGIB, and we particularly focused on re-prescribing without PPI coverage. The data originated from a previous study on drug-related UGIB by this group.[6] Our focus was on accounting for actual prescribing behaviour rather than a normative analysis of guideline adherence.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Setting

The data in this study were retrieved from three different sources, a recent study on drug-related UGIB, Odense University Pharmaco-epidemiological Database (OPED) and The National Board of Health in Denmark.

Our group has previously conducted a population-based case-control study where we reported an association between SSRI-use and UGIB.[6] The case material was defined in this previous publication.

In brief, serious UGIB cases were identified by a manual review of all discharge summaries from hospitalisations with the diagnosis of peptic ulcer (PU) as the main diagnosis (complicated or not) or gastritis (ICD10-code K25-29) during 1 August 1995 to 31 July 2006. A total of 3652 cases were found that fulfilled all of the following criteria:

  1. Admission with PU or gastritis as main diagnosis within one of the county's hospitals during the period 1 August 1995 to 31 July 2006.
  2. Significant bleeding defined by either melaena, subnormal haemoglobin or need for transfusions.
  3. A potential bleeding source in the stomach or duodenum identified by endoscopy or surgery. Variceal bleedings were excluded.

Characteristics of the cases are detailed in a previous publication.[6]

Through linkage with OPED we identified 1890 cases that were users of one or more drugs with antihaemostatic action (DAHA) at the time of admission. Combined users contribute to both survival analyses in the present article.

Since 1990 information on reimbursed drug dispensing in the County of Funen has been recorded in the OPED and from 1 January 2007 the entire Region of Southern Denmark (population 1.2 m). Each prescription record includes a person identifier, the date of dispensing, the brand, quantity and form of the drug. The substances and quantities are registered according to WHO's anatomical-therapeutic-chemical (ATC) system and defined daily doses (DDD) methodology.[7] Information on indication for treatment and dosing instructions are not recorded. Over-the-counter (OTC) drugs and some prescription drugs, mainly oral contraceptives, hypnotics, sedatives and some antibiotics[8] are not reimbursed and therefore not recorded in the database.

Information on deaths and migrations was retrieved from the registers in the National Board of Health.

Data were linked using the mutual person identifier, the Central Person Registry (CPR)-code, which is shared with virtually all other health related registries in Denmark, thereby allowing record-linkage studies.[9]

An ethics committee approval was not required. The study was approved by the Danish Data Protection Agency.

Analysis

We chose to study re-prescribing of the following DAHA; NSAID, low-dose ASA, SSRIs, VKA, clopidogrel and dipyridamol. These drugs are known to be associated with UGIB when used in a single drug regime or in combination with other DAHA.[6, 10, 11] For a description of how we defined exposure to these drugs, we refer to the original publication.[6] Although SSRIs are usually not considered as antiplatelet drugs, they have similar pharmacological actions[12] and have been shown to be associated with UGIB.[6] We therefore decided to include them in the analysis.

We analysed the re-prescribing of the same DAHA that the patients had been taking prior to the UGIB. That is in cases that were exposed to NSAID at their UGIB episode, we studied the re-prescribing of NSAID after discharge. In this respect, we allowed that the re-introduced NSAID could be different from the NSAID prescribed before the UGIB. Similar considerations were employed for VKA and SSRIs. We used the Kaplan–Meier method to estimate the time from UGIB to re-prescribing. Patients were formally considered unexposed until they redeemed their first prescription on the DAHA in question. Follow-up began on the date of discharge from the UGIB episode, and subjects were censored from the analysis in case of the following events: death, emigration from the source population, end of 10 years’ observation or 1 January 2009, whichever came first. In this part of the analysis, we did not take use of PPIs into account.

Re-prescribing without a PPI

We subsequently generated Kaplan–Meier curves of the time from UGIB to the first prescription of the DAHA without concurrent use of PPIs. For each prescription of a PPI we defined the period of use as follows: the period started on the date of redeeming the prescription and lasted a number of days equivalent to the number of DDDs dispensed times two. That is we assumed a daily intake of PPIs of 0.5 DDD/day. If a DAHA prescription was redeemed on a date that was covered by a PPI prescription as defined above, this DAHA was considered as taken with concurrent PPI, and follow-up continued until the first occurrence of the DAHA without PPI coverage.

To fully interpret the analysis of DAHA re-prescribing without PPI, we also analysed the pattern of use of PPIs after discharge for patients who have not taken DAHA prior to their UGIB. For these subjects, we calculated the proportion that would have been deemed exposed to PPIs by the definition above on the beginning of each 30-day-period after discharge; i.e., on day 30, day 60, day 90 etc. Subjects were excluded from this analysis when they died, emigrated or if they reached the end of the study period Figure 2.

All analyses were performed using stata release 10.[13]

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The baseline characteristics of the DAHA subjects are shown in Table 1. Among the 3652 cases of UGIB; 377 (10.3%), 1213 (33.2%), 621 (17.0%), 164 (4.5%), 56 (1.5%) and 144 (3.9%) were users of SSRI, NSAID, low-dose ASA, VKA, clopidogrel and dipyridamol at the time of admission respectively. One year after the UGIB, 82% of SSRI users, 25% for NSAID users, 43% of low-dose ASA users, 68% of VKA users, 55% of clopidogrel users and 71% of dipyridamol users had redeemed a new prescription Table 2. A total of 1762 patients were non-users of any of the above DAHA prior to their UGIB and were not included in the analysis of re-prescribing.

Table 1. Characteristics of 1890 subjects using drugs with an antihaemostatic action
  1. Unless otherwise indicated, data are shown as numbers. PPI, proton pump inhibitor; SSRI, selective serotonin reuptake inhibitor; H2RA, H2-receptor antagonist; HP, Helicobacter pylori; s.d., standard deviation; GERD, gastroesophageal reflux disease.

  2. a

    ICD-8 295-300; ICD-10 F20, 30–33, K70.0-K70.9, and F10.

Age, mean (s.d.)75.1 (12.5)
<55182 (9.6%)
≥55 & <74559 (29.6%)
≥751149 (60.8%)
Men854 (45.2%)
Current drug use (ATC code)
Low-dose aspirin (B01AC06)621 (32.9%)
Anticoagulants (B01AA)164 (8.7%)
NSAID (M01A)1213 (64.2%)
SSRIs (N06AB)377 (19.9%)
Clopidogrel (B01AC04)56 (2.9%)
Dipyridamol (B01AC07)144 (7.6%)
Systemic corticosteroids (H02AB)226 (12.0%)
H2RA (A02BA)131 (6.9%)
PPIs (A02BC)269 (14.2%)
Statins (C10AA)164 (8.7%)
Nitrates (C01DA)203 (10.7%)
TCA (N06AA)51 (2.7%)
Spironolactone (C03DA01)118 (6.2%)
History of
Hp eradication76 (4.0%)
Chronic obstructive pulmonary disease165 (8.7%)
Peptic ulcer162 (8.6%)
Upper gastrointestinal bleeding36 (1.9%)
Ischaemic heart disease408 (21.6%)
Hepatic cirrhosis18 (1.0%)
Diabetes mellitus163 (8.6%)
Renal failure41 (2.2%)
Heart failure261 (13.8%)
Hypertension331 (17.5%)
Stroke233 (12.3%)
Alcohol-related diagnosis a92 (4.9%)
Alcohol-related prescriptions (disulfiram)29 (1.5%)
Alcohol-related diagnosis or drug use111 (5.9%)
GERD74 (3.92%)
Table 2. Characteristics of cases using drugs with antihaemostatic action and their cumulative re-prescribing rates of the relevant drugs used prior to the upper gastrointestinal bleeding during a 5 years span. The re-prescribing rates are reported with 95% confidence intervals
DrugsThe sum of users of the following potential causative drugs; (UGIB = 3652)Mean ageMen (%)3 months6 months12 months36 months60 months
SSRI377 (10.3%)75.8136 (36.1)0.60 (0.55–0.65)0.72 (0.67–0.77)0.82 (0.77–0.86)0.89 (0.85–0.92)0.91 (0.87–0.94)
NSAID1213 (33.2%)74.7507 (41.8)0.11 (0.09–0.13)0.18 (0.15–0.20)0.25 (0.23–0.28)0.44 (0.41–0.48)0.57 (0.53–0.61)
Low-dose ASA621 (17.0%)77.9285 (45.9)0.20 (0.17–0.24)0.33 (0.29–0.37)0.43 (0.39–0.48)0.55 (0.50–0.59)0.60 (0.55–0.65)
Vitamin K antagonist164 (4.5%)74.9103 (62.8)0.41 (0.33–0.49)0.55 (0.47–0.63)0.68 (0.60–0.75)0.74 (0.65–0.82)0.74 (0.65–0.82)
Clopidogrel56 (1.5%)71.730 (53.6)0.40 (0.28–0.55)0.46 (0.34–0.61)0.55 (0.42–0.69)0.65 (0.51–0.79)0.65 (0.51–0.79)
Dipyridamol144 (3.9%)76.580 (55.6)0.55 (0.46–0.63)0.61 (0.52–0.69)0.71 (0.62–0.80)0.74 (0.65–0.82)0.74 (0.65–0.82)

Table 3 describes the re-prescribing pattern for DAHA used without a PPI. The cumulative rate after 1 year was 25% of SSRI users, 3% for NSAID users, 5% for low-dose ASA users, 1% for VKA users, 17% for clopidogrel users and 7% for dipyridamol users.

Table 3. Cases’ cumulative re-prescribing rates of drugs with antihaemostatic action (DAHA) used prior to the upper gastrointestinal bleeding during a 5 years span. The specific DAHA are re-prescribed without additional proton pump inhibitors. The re-prescribing rates are reported with 95% confidence intervals
 3 months6 months12 months36 months60 months
SSRI0.16 (0.12–0.20)0.21 (0.17–0.26)0.25 (0.20–0.30)0.37 (0.32–0.44)0.43 (0.36–0.51)
NSAID0.01 (0.01–0.02)0.02 (0.01–0.03)0.03 (0.02–0.05)0.08 (0.06–0.10)0.11 (0.09–0.14)
Low-dose ASA0.01 (0.01–0.03)0.02 (0.01–0.04)0.05 (0.03–0.07)0.07 (0.05–0.10)0.10 (0.07–0.14)
Vitamin K antagonist0.00 (–)0.01 (0.00–0.05)0.01 (0.00–0.05)0.02 (0.00–0.07)0.02 (0.00–0.07)
Clopidogrel0.13 (0.07–0.26)0.15 (0.08–0.28)0.17 (0.09–0.31)0.21 (0.12–0.36)0.21 (0.12–0.36)
Dipyridamol0.01 (0.00–0.06)0.01 (0.00–0.06)0.06 (0.03–0.12)0.07 (0.03–0.14)0.07 (0.03–0.14)

The findings in Tables 2 and 3 are illustrated in the Kaplan–Meier graphs in Figure 1. The upper curve (dashed line) in each Kaplan–Meier curve illustrates the time until a DAHA is redeemed, irrespective of PPI use. The lower curve (solid line) represents the time until re-prescribing of a DAHA without a PPI.

image

Figure 1. Kaplan–Meier analysis illustrating the time span before a new prescription of NSAID, SSRI, low-dose ASA, vitamin K antagonist, dipyridamol or clopidogrel is redeemed after an episode of serious upper gastrointestinal bleeding. The solid line (lower curve) represents the time until re-prescribing of a drug with an antihaemostatic action (DAHA) without a PPI. The dashed line (upper curve) illustrates the time until a DAHA is redeemed, irrespective of PPI use.

Download figure to PowerPoint

image

Figure 2. Proton pump inhibitors use among 1762 patients with an upper gastrointestinal bleeding that were all non-users of SSRI, low-dose ASA, NSAID, warfarin, clopidogrel and dipyridamol.

Download figure to PowerPoint

Figure 2 illustrates the use of PPI among 1762 patients who were all non-users of DAHA prior to their UGIB event. Initially, the use of PPI was 67%, but stabilised on approximately 30% after a few months.

We conducted a subanalysis to explore for any changes in prescription trends over the study period. We divided our data material into two groups (1 August 1995 to 31 December 1999 and 1 January 2000 to 31 July 2006). No significant changes were found (data not shown).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We found that users of SSRI, VKA, clopidogrel and dipyridamol had a relatively high cumulative rate of re-prescribing (55–82%) within the first year and that the prescriptions were redeemed shortly after discharge from hospital. This is an elderly population and use of DAHAs may put them in a very high risk of re-bleeding. For users of NSAID and low-dose ASA, we found a more slow reintroduction. The rate of re-prescribing without PPIs was markedly lower for all drug classes. Particularly low rates were seen for NSAID, low-dose ASA and VKA.

There are two likely explanations for the marked difference between Kaplan–Meier curves with and without PPIs taken into consideration: the PPIs are prescribed simply because of a recent serious bleeding episode, or the PPIs are prescribed specifically to prevent UGIB from the re-prescribed DAHA. The use of PPIs among subjects who had not taken DAHA prior to their UGIB Figure 2 suggests the latter. The proportion of PPI use was as low as 30%, which would only explain a small difference between the Kaplan–Meier curves with and without PPI taken into consideration.

The curves are determined by several other factors: the physicians’ or patients’ fear of new UGIBs, their awareness of the UGIB risk with different DAHAs and the strength of the indication for the DAHA. For example, the slow reintroduction of NSAID and its high coverage suggest a high degree of awareness of NSAID associated bleeding and a non-imperative indication, i.e. that the NSAID is not deemed vital for the patient. The relatively high re-prescribing rate and low PPI coverage for SSRIs might reflect a low awareness of the SSRI-UGIB association. The fast re-introduction for VKA and strikingly high PPI coverage suggest an imperative indication and a high awareness of the UGIB risk. These patterns may not necessarily reflect the current evidence. NSAID are by far the drug class with the strongest association with UGIB.[14] VKA and ASA are not particularly prone to induce UGIB, having ORs in the order of two.[10] PPI use has been documented as a prophylactic measure against UGIB related to NSAID,[5] low-dose ASA,[5] clopidogrel[5, 15, 16] and SSRI[6] use, but to a lesser degree for VKA[3] and dipyridamol. We find it reassuring that low-dose ASA is quickly reintroduced in the vast majority of patients. A newly published controversial study by Sung et al. suggests that prophylactic low-dose ASA should be continued even during the current bleeding event. The increased risk for recurrent UGIB is outweighed by a reduced overall mortality.[17]

We have only found few other studies investigating the re-prescription pattern of DAHA after an episode of UGIB. Massó González et al. have reported data on the risk of recurrent UGIB associated with continued use of potential causative drugs.[3] Other groups have reported that PPI utilisation is suboptimal in certain risk groups using low-dose ASA and NSAID.[18, 19] Even though our results are only indirectly comparable to these studies, our results imply that physicians generally are very cautious and to a high extent prescribe PPI to patients using DAHA.

Massó González et al. found that 21% of long-term users were still using NSAID 1 year after an episode of UGIB and in the group of long-term low-dose ASA users, 28% were using low-dose ASA 1 year after the bleeding event.[3]

The current study has several strengths; first we used only manually validated cases from a previous study by this group.[6] Second; the study is a true population-based approach with full coverage of all admissions and prescription data. A potential limitation to this study is, as in any observational study, we cannot be certain of the patients’ compliance. Also, some NSAID and low-dose ASA was available over-the-counter during the study period. However, the coverage of our data source for these drugs, i.e., the proportion sold on prescription is 88% and 98.5% respectively.[10] Another potential limitation could be the low odds ratios associating VKA, ASA, SSRIs, clopidogrel and dipyridamol with UGIB. Most are in the order of 2, which imply an attributable proportion of 0.5. In other words only half of the UGIB patients who were exposed to these drugs were associated with them. The other half was in reality bleedings that would have occurred irrespective of DAHA use. As we can never know which half is drug induced, it is difficult to use this in clinical management.

In conclusion, we find a markedly rational approach to re-prescribing of DAHA in patients who have experienced a serious adverse reaction. Particularly, the very extensive use of PPIs in these patients is reassuring and emphasises the clinicians’ awareness of the issue.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Declaration of personal interests: JH has received research grants from Novartis, MSD, Nycomed and Pfizer, and has taught on courses for the Association of the Danish Medicines Industry. All other authors have no conflicts of interest. Declaration of funding interests: The University of Southern Denmark and Funen County provided data free of charge.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Rockall TA, Logan RF, Devlin HB, et al. Incidence of and mortality from acute upper gastrointestinal haemorrhage in the United Kingdom. Steering Committee and members of the National Audit of Acute Upper Gastrointestinal Haemorrhage. BMJ 1995; 311: 2226.
  • 2
    Kapur KC, Green JT, Turner RG, et al. Auditing mortality from upper gastrointestinal haemorrhage: impact of a high dependency unit. J R Coll Physicians Lond 1998; 32: 24650.
  • 3
    Massó González EL, García Rodríguez LA. Proton pump inhibitors reduce the long-term risk of recurrent upper gastrointestinal bleeding: an observational study. Aliment Pharmacol Ther 2008; 28: 62937.
  • 4
    Bhatt DL, Scheiman J, Abraham NS, et al. ACCF/ACG/AHA 2008 expert consensus document on reducing the gastrointestinal risks of antiplatelet therapy and NSAID use. Am J Gastroenterol 2008; 103: 2890907.
    Direct Link:
  • 5
    Lanas A, García-Rodríguez LA, Arroyo MT, et al. Effect of antisecretory drugs and nitrates on the risk of ulcer bleeding associated with nonsteroidal anti-inflammatory drugs, antiplatelet agents, and anticoagulants. Am J Gastroenterol 2007; 102: 50715.
  • 6
    Dall M, Schaffalitzky de Muckadell OB, Lassen AT, et al. An association between selective serotonin reuptake inhibitor use and serious upper gastrointestinal bleeding. Clin Gastroenterol Hepatol 2009; 7: 131421.
  • 7
    WHO Collaborating Centre for Drug Statistics Methodology. ATC index with DDDs and guidelines for ATC classification and DDD assignment. Oslo: Norwegian Institute of Public Health, 2006.
  • 8
    Gaist D, Sørensen HT, Hallas J. The Danish Prescription Registries. Dan Med Bull 1997; 44: 4458.
  • 9
    Frank L. Epidemiology. When an entire country is a cohort. Science 2000; 287: 23989.
  • 10
    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.
  • 11
    Johnsen SP, Sørensen HT, Mellemkjoer L, et al. Hospitalisation for upper gastrointestinal bleeding associated with use of oral anticoagulants. Thromb Haemost 2001; 86: 5638.
  • 12
    de Abajo FJ, Montero D, Rodríguez LA, et al. Antidepressants and risk of upper gastrointestinal bleeding. Basic Clin Pharmacol Toxicol 2006; 98: 30410.
  • 13
    Release 10.1. StataCorp, 4905 Lakeway Drive, College Station, TX, USA. Available at: http://www.stata.com.
  • 14
    Hernández-Díaz S, Rodríguez LA. Association between nonsteroidal anti-inflammatory drugs and upper gastrointestinal tract bleeding/perforation: an overview of epidemiologic studies published in the 1990s. Arch Intern Med 2000; 160: 20939.
  • 15
    Chin MW, Yong G, Bulsara MK, et al. Predictive and protective factors associated with upper gastrointestinal bleeding after percutaneous coronary intervention: a case-control study. Am J Gastroenterol 2007; 102: 24116.
    Direct Link:
  • 16
    Ibáñez L, Vidal X, Vendrell L, et al. Upper gastrointestinal bleeding associated with antiplatelet drugs. Aliment Pharmacol Ther 2006; 23: 23542.
  • 17
    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: 19.
  • 18
    Coté GA, Norvell JP, Rice JP, et al. Use of gastroprotection in patients discharged from hospital on nonsteroidal anti-inflammatory drugs. Am J Ther 2008; 15: 4449.
  • 19
    Targownik LE, Al-Mamfud A. The prevalence of risk factors for gastrointestinal complications and use of gastroprotection among persons hospitalized for cardiovascular disease. Aliment Pharmacol Ther 2006; 23: 7439.