Use of gastroprotective agents in recommended doses in hospitalized patients receiving NSAIDs: a drug utilization study
- First Online:
- Cite this article as:
- Erdeljic, V., Francetic, I., Macolic Sarinic, V. et al. Pharm World Sci (2006) 28: 318. doi:10.1007/s11096-006-9062-2
- 62 Views
In recent years, studies investigated to what extend recommendations for co-prescribing gastroprotective agents in prevention of NSAID-induced gastrointestinal complications are followed in clinical practice. However, only a few studies have also taken into consideration the recommended dose of gastroprotectives prescribed in NSAID-induced ulcer prophylaxis. The aim of our study was to evaluate the prevalence of concomitant use of gastroprotectives with NSAIDs in hospitalized patients, with emphasis on the recommended dose of gastroprotectives for ulcer prophylaxis.
This observational, cross-sectional, drug utilization study included all adult patients receiving NSAIDs hospitalized in the Clinical Hospital Center Zagreb on the day of the study. Data on age, sex, comorbidities, indications for NSAID use, type/dose of NSAIDs and gastroprotectives, history of gastrointestinal events, active gastrointestinal symptoms and risk factors were evaluated.
Main outcome measure
Study outcomes were: (1) prevalence of prescription of gastroprotectives among NSAID-users at risk; (2) prevalence of prescription of gastroprotective in recommended dose; (3) association between risk factors and prescription of GPAs.
The rates of gastroprotectives prescription were significantly higher in NSAID-users with concomitant risk factors as compared to patients without risk factors [47/70 (67.1%) and 8/22 (36.4%), respectively; p = 0.01072]. However, gastroprotection in recommended ulcer-preventive dose was low in both groups [8/70 (11.4%) and 9/92 (9.8%), respectively]. The number of concomitant risk factors did not increase the odds of receiving anti-ulcer therapy (odds ratio 0.7279). Thirty-three percent of patients with concomitant risk factors were not prescribed gastroprotectives. Ibuprofen, NSAID with the lowest risk of inducing gastrointestinal complications, was prescribed in only two patients.
The results indicate high awareness among hospital physicians about possible NSAID-induced gastrointestinal complications, but insufficient knowledge about risk factors related to NSAID-induced gastrointestinal toxicity, recommended dose of gastroprotectives in NSAID-induced ulcer prophylaxis and gastrointestinal toxicity of different types of NSAIDs.
KeywordsAnti-ulcer agents CroatiaDrug utilizationGastroprotectionNon-steroidal anti-inflammatory agentsPreventionRisk factorsSecondary care
Impact on practice
The prevalence of gastroprotection in our study was relatively high, but when recommended daily dose of gastroprotective agents taken into consideration, the actual prevalence of gastroprotection was unacceptably low. Similar drug use patterns may be expected elsewhere.
The knowledge about gastrointestinal toxity and effective gastroprotection could be improved in Croatia.
Further studies with the aim to assess the use of gastroprotective agents with NSAIDs in secondary care are needed as well as continuous education.
Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most commonly used medications . A major factor limiting their use is gastrointestinal (GI) toxicity, ranging from dyspepsia to life-threatening events. It is estimated that up to 15–30% of patients taking NSAIDs develop GI adverse effects, and that significant GI events occur in 2–4% patients taking NSAIDs .
NSAIDs are exhibiting their effect by inhibition of cyclooxygenase (COX-1 and COX-2) enzymes. Non-selective NSAIDs inhibit both COX-1 and COX-2 to varying degrees, where the anti-inflammatory effect is derived from inhibition of COX-2, while the adverse effects arise from inhibition of COX-1 activity. Although it was initially anticipated that COX-2 specific inhibitors would be safer than non-specific NSAIDs [3, 4], cardiovascular complications now limit their use . The recent safety concerns surrounding the entire class of COX-A inhibitors [6, 7], are resulting in the reestablishment of conventional, nonselective NSAIDs as a mainstay of clinical care for patients with musculoskeletal disorders.
Factors associated with increased risk of NSAID-associated serious complications are history of ulcer complications, concomitant anticoagulant therapy, advanced age (>65 years), concomitant corticosteroid use, chronic major organ impairment, the use of high dose or multiple NSAIDs and severe rheumatoid arthritis [8–10]. Strategies in prevention of NSAID-induced GI events are: (1) acetaminophen as the first-line therapy in musculoskeletal disorders ; (2) use of less GI toxic NSAIDs; (3) use of the lowest effective dose of NSAID; (4) concomitant use of gastroprotective agents (GPAs) in patients with increased risk.
Prevention of NSAID-induced GI morbidity by co-prescription of GPAs has been validated in many clinical studies [12–14]. The use of GPAs has focused on two approaches: prostaglandin replacement (misoprostol) and inhibition of acid secretion (proton pump inhibitors and histamine2-receptor antagonists). Misoprostol was of no relevance for our study since it is not approved in Croatia. Histamine2-receptor antagonists (H2RA) heal almost all NSAID ulcers when the patient discontinues NSAID use. However, in patients who continue NSAID use, H2RA in traditional doses are more effective in healing duodenal ulcers than gastric ulcers [15, 16]. It appears that larger than traditional doses of H2RAs are more effective at NSAID-associated ulcer prevention . Several studies have confirmed the superior efficacy of proton pump inhibitors (PPIs) in the short and longer-term prevention of NSAID-induced ulcers as compared to H2RA [14, 18]. However, prophylactic use of PPIs in all patients is unnecessary and cost-prohibitive .
Two recent systematic reviews have assessed the effectiveness of various GPAs against GI complications induced by NSAIDs use. A Cochrane review by Rostom et al. assessed the effectiveness of H2RAs, PPIs and misoprostol against endoscopic ulcers, ulcer complications, symptoms and dropouts . The results showed that all GPAs included prevented endoscopic ulcers. A systematic review by Hooper et al. assessed the effectiveness of H2RAs, PPIs and misoprostol plus non-selective NSAIDs, and COX-2 selective NSAIDs in reducing serious GI complications, symptomatic ulcers, serious cardiovascular/renal disease, death, and improving quality of life. They found that misoprostol, COX-2 specific NSAIDs and probably PPIs significantly reduce the risk of symptomatic ulcers, and misoprostol and probably COX-2 selective NSAIDs significantly reduce the risk of serious GI complications. Data were insufficient to draw conclusions on effect of H2RAs compared with placebo on any primary outcomes. Endoscopic ulcers were significantly reduced in participants taking H2RAs compared to placebo. In half of the 15 studies included the dose of H2RAs was higher than the traditional dose . These reviews support the consensus that PPIs and probably high doses of H2RAs offer significant protection in patients receiving NSAIDs.
The prevalence of co-prescription of GPAs with NSAIDs is estimated to be 20–40% [22, 23]. According to recommendations from professional societies such as The American College of Rheumatology [24, 25], patients with at least one gastrointestinal risk factor should receive NSAID plus a co-prescribed protective agent. Several countries have established relevant clinical guidelines. However, the extent to which such guidelines are implemented in clinical practice is not known. There are only a few studies that have also taken into consideration the recommended dose of H2RAs when used in NSAID-ulcer prophylaxis [17, 26].
Rational and evidence based drug prescribing is one of the main goals in pharmaceutical care. Underutilization, as well as inappropriate use of GPAs, can contribute to increased health costs . It is estimated that anti-ulcer drug prescriptions consume around 10% of the national drug budgets in Australia and England [28–30], but not all anti-ulcer drug prescribing is associated with NSAIDs. Only by optimal prescribing it is possible to control the rising share of these drugs in national drug budgets , as well the costs of NSAIDs related adverse events. Although rational NSAIDs related GPAs prescribing is likely to increase expenditure in this group of medicines, this increase could be nevertheless estimated as less as compared to the costs of NSAIDs related adverse event.
The purpose of this study was to evaluate not only the prevalence of concomitant use of GPAs with NSAIDs, but also the extent to which NSAID users receive therapy that really provides gastroprotection. Previous studies assessed the utilization of gastroprotective measures in primary care. The aim of our study was to assess the use of GPAs with NSAIDs in secondary care. The fact that most GI complications induced by NSAID therapy are occurring during the start of NSAID therapy makes the secondary care prescribing even more important. Although most prescribing takes place in general practice, the influence of secondary care prescribing on primary care prescribing is well recognized [32–35].
We conducted an observational, cross-sectional study on November 11th, 2003. The study included all adult patients (aged > 18 years) hospitalized in the Clinical Hospital Center Zagreb on the day of the study who received NSAIDs. An exclusion criterion was an indication for which gastroprotective drugs were prescribed other than NSAID ulcer prophylaxis. A structured questionnaire was developed at the Division of Clinical Pharmacology with the aim to collect information on age, sex, reason for hospitalization, concomitant diseases, indications for use of NSAIDs or/and GPAs, type and dose of NSAIDs and GPAs prescribed, initiation of treatment with NSAIDs or/and GPAs (during hospitalization/before hospitalization), history of GI disorders, presence of active GI symptoms and concomitant risk factors. On the day of the study, 10 clinical pharmacologists and residents in clinical pharmacology visited all departments and collected data on all patients prescribed NSAIDs according to the prepared questionnaire. Informed oral consent was obtained from all patients prior to inclusion. The data collection was anonymous. Included were patients prescribed gastroprotectives as NSAID ulcer prophylaxis, which was determined by interview and/or reviewing clinical charts. Use of recommended gastroprotective strategy was defined as the co-prescription of PPIs in standard dose or H2RAs in double dose. The co-prescription of an antacid or H2RA in standard dose was considered as less than recommended in NSAID ulcer prophylaxis.
At the time of the study, there were no local guidelines on co-prescription of GPAs with NSAIDs or on indications/contraindications for the use of NSAIDs. The prescription of NSAIDs as well as eventual co-prescription of GPAs was left to the knowledge of the prescribing physician. In evaluating the appropriateness of co-prescription of GPAs with NSAIDs, we used the guidelines of the American College of Rheumatology [24, 25].
Prescribed GPAs included antacids (Al-hydroxide–Mg-carbonate gel), H2RA ranitidine and PPIs (omeprazole, pantoprazole, lansoprazole). The recommended daily doses of GPAs were defined as ranitidine 600 mg, omeprazole 20 mg, pantoprazole 40 mg and lansoprazole 30 mg. Lower doses of these drugs were classified as “less than recommended”. Antacids were considered inappropriate gastroprotective strategy in prevention of NSAID-induced GI adverse effects. Misoprostol was not approved in Croatia at the time of the study.
Identification of patients at risk
Factors that identified patients at high risk of developing NSAID associated complications were defined as age >65 years, history of GI events, present GI symptoms (dyspepsia, gastritis), concomitant use of corticosteroids, anticoagulant therapy or of aspirin, serious co-morbidities, ≥2 prescribed NSAID, higher than recommended daily doses of NSAIDs. In patients with 1≥risk factors the co-prescription of GPAs was considered recommended. Standard daily doses of NSAIDs were defined as ketoprofen 300 mg, diclofenac 100 mg, indomethacin 50 mg, ibuprofen 1200 mg and piroxicam 20 mg.
Study outcomes were: (1) prevalence of GPAs prescription among patients at risk for NSAID-induced GI complications vs. patients not at risk; (2) prevalence of GPAs prescription in recommended doses; (3) estimation of the association between risk factors and prescription of GPAs.
Standard descriptive statistics were used to describe the study population and their utilization of NSAIDs and GPAs. Means and Standard Deviations (SD) were calculated for age. The proportions of NSAID users with and without gastroprotective therapy in relation to concomitant risk factors and type of GPAs prescribed were calculated and compared using the χ2-test. The relationship between co-prescription of GPAs and number of concomitant risk factors by estimation of odds ratio, was performed using the logistic regression model. All tabulations and statistical analysis were done using Statistica for Windows, Version 5.5, StatSoft, Inc. (2000).
Out of 770 patients hospitalized on the day of the study, 93 patients received NSAIDs. One patient with rheumatoid arthritis was prescribed a COX-2 selective NSAID (rofecoxib). This patient was excluded being the only one receiving a selective COX-2 inhibitor.
Patients taking NSAIDsa
Age, mean ± SD
58.9 ± 13.5
Age > 65
NSAIDs prescribed during hospitalizationb
GPAs prescribed during hospitalizationc
No risk factor
1 risk factor
2 risk factors
3 risk factors
4 risk factors
NSAIDs prescription by risk factors
≥2 risk factors, N
1 risk factor, N
No risk factors, N
Diclofenac + indomethacin
Ketoprofen + indomethacin
Diclofenac + ketoprofen
GPAs prescription by risk factors
Number of patients
Absent of GPA [N (%)]
Less than recommended dose [N (%)]
Recommended dose [N (%)]
No risk factors
Specific risk factors
Age > 65 years
History of GI disorders
History of peptic ulcers
Presence of GI symptoms
Concomitant corticosteroids therapy
Concomitant aspirin therapy
Any risk factor
p = 0.01072
≥2 risk factors
1 risk factor
2 risk factors
3 risk factors
4 risk factors
Prescription of gastroprotective agents by type and dose recommendation
Number of patients
No GPAs [N (%)]
Less than recommended dose [N (%)]
Recommended dose [N (%)]
Total anti-ulcer therapy
Anti-ulcer therapy in patients at risk
Although the prevalence of prescribing GPAs overall and in NSAID users at high risk was relatively high (59.8 and 67.1%) compared to similar studies [36–39], when recommended daily doses of GPAs were taken into consideration, the actual prevalence of gastroprotection was unacceptably low (9.8 and 11.4%, respectively). High prevalence of GPA utilization with NSAIDs may be contributed to the high prevalence of concomitant risk factors and higher awareness of possible NSAID-induced GI effects among hospital physicians . Higher prevalence of individual risk factors as compared to other studies could be contributed to the more vulnerable hospital population (more co-morbidities, older age, concomitant therapy). Furthermore, the hospitalization itself could be considered a risk factor, but was not taken into account as a separate risk factor in this study. Although the presence of any risk factor among NSAID users resulted in significantly higher prevalence of GPA prescription, the prescription rates of GPAs were similar across all risk groups. The number of concomitant risk factors did not increase the odds of GPA prescription. The high prevalence of GPA prescription without significant difference in prescription rates between different risk groups could be explained by awareness of possible NSAID-induced GI complications, but insufficient knowledge about individual risk factors related to increased risk. Here is where clinical pharmacy would have a significant role in identifying patients at risk of NSAIDs induced GI adverse events and recommending appropriate prophylaxis as well as identifying those who are unnecessarily receiving prophylaxis. Unfortunately, clinical pharmacy in Croatia is at its beginning, so there are no clinical pharmacists at our institutions.
The most prescribed NSAID was ketoprofen, which is one of the NSAIDs with highest relative risk of inducing GI events. Ibuprofen, the NSAID with the lowest relative risk of inducing GI events , was prescribed in only two patients. The high prescription rate of ketoprofen is almost entirely contributed to the Department of Surgery where ketoprofen was the exclusively prescribed NSAID. The data collected did not include the reasons for the choice of a specific NSAID, so the high prescription rate of ketoprofen remains open to discussion. One patient with present duodenal ulcer (confirmed by endoscopy) was receiving ketoprofen with co-prescription of a PPI. The prescription patterns of NSAIDs demonstrate lack of knowledge about differences in GI toxicity of various types of NSAIDs.
All patients receiving PPIs were appropriately protected, since PPIs are effective in ulcer prophylaxis in their standard dose. H2RAs have to be taken in double dose to be effective in ulcer prophylaxis . Although PPIs are considered the appropriate gastroprotective drug in patients with high risk for NSAID-induced GI toxicity, due to the perception of their lower cost, in our hospital H2RAs are still the most prescribed GPAs (45/55; 81.8%). The costs of daily ulcer preventive dose of ranitidine, omeprazole, lansoprazole and pantoprazole is € 1.09, € 0.96, € 0.95 and € 1.02, respectively. Only one patient received H2RAs in ulcer preventive dose. This indicates insufficient knowledge about dose recommendations of H2RAs when used in NSAID-ulcer prophylaxis, as well as about the costs of daily therapy with H2RAs and PPIs in ulcer preventive doses. Antacids are not considered appropriate therapy in NSAID-induced ulcer prophylaxis. However, three patients received antacids in ulcer prevention, one patient in combination with ranitidine.
Monitoring of GPA prescription strategies in NSAID users in hospitals is important due to their influence on prescribing habits in general practice [32–35], where the GPA prescription rate is low, ranging from 7, 9–41, 6% [41, 42]. The risk for serious gastrointestinal complications seems to be related to the beginning of NSAID use , and according to our results more then two thirds of patients were prescribed NSAIDs and GPAs during hospitalization. However, regarding the influence on primary health care prescribing, it would be useful to further differentiate information on acute vs. chronic use of NSAIDs and to assess the extent of repeat prescribing of NSAIDs and GPAs in primary care.
The methodological weaknesses of our study additionally include lack of data on duration of NSAID therapy, long/short term prescription, concomitant drug therapy, as well as no categorization by indications for NSAID use (analgesia or rheumatic disease), as well as no assessment of appropriateness of the NSAID prescription in the individual patient. The evaluation of side effects of NSAIDs use during hospitalization would further help to precisely evaluate the impact of the quality of NSAIDs/GPAs use.
The results of this study pointed out at the need for local guidelines, continuous evaluation and educational measures. Based on these results we decided to make this study a first step in an interventional study. The planned interventions included: the presentation of the study results at a monthly clinical staff meeting at our hospital; publication of guidelines in the hospital bulleting [guidelines for NSAIDs and GPAs prescription (related and unrelated to NSAIDs use), information on GI toxicity of different NSAIDs, information on efficacy and appropriate dose of different GPAs in various indications, information on costs of daily therapy of GPAs in ulcer preventive doses]; publication of guidelines in a form of posters distributed to every Department; consultations with clinical pharmacologists. The study design will be improved in regard to weaknesses listed above. In response to this study, we further conducted a study aimed at the assessment of quality of pain management at the Department of Medicine, which was followed by publication of local guidelines for pain management in the hospital bulleting.
Despite its weaknesses, we believe that our study offers useful information on GPAs prescription strategies in NSAID users in the hospital setting, with the emphasis on the extent to which NSAID users receive recommended therapy that really provides gastroprotection. More studies assessing the use of H2RAs as ulcer-preventive drugs are needed, as well as studies dealing with the evaluation of GPAs use with NSAIDs in secondary care.
Although we did not directly measure the knowledge of hospital physicians on the studied topic, the results of our study nevertheless indicate high awareness of possible NSAID induced GI complications among hospital physicians (high GPA prescription rate), but also insufficient knowledge about specific risk factors related to NSAID-induced toxicity, the costs of, and recommended dose of H2RAs in ulcer prophylaxis and GI toxicity of different types of NSAIDs. Although the prevalence of GPAs among NSAID users overall and in patients at high risk was relatively high, when recommended daily doses of GPAs were taken into consideration, the actual prevalence of gastroprotection was unacceptably low. This indicates the need for continuous educational measures. More than 2/3 of patients receiving NSAIDs were prescribed NSAIDs and GPAs for the first time during hospitalization. Since secondary care prescribing may have considerable influence on general practicioners’ prescribing, studies aimed at evaluation of the use of GPAs with NSAIDs in secondary care are needed. In conclusion, it should be emphasized that drug utilization studies should be promptly followed by educational and other appropriate measures for quality improvement for the benefit of our patients.
Potential conflicts of interest: none declared.