Eur J Clin Pharmacol (2000) 56: 739±746 DOI 10.1007/s002280000221
P H A R M A C O E PI D E M I O L O G Y A N D PR E S C R I P T I O N
R. Fijn á A. G. G. Stuurman-Bieze á P. B. van den Berg J. R. B. J. Brouwers á P. A. de Grae L. T. W. de Jong-van den Berg
Predictors for prophylactic antithrombotic prescribing in ischaemic heart disease and the impact of national guidelines Received: 24 May 2000 / Accepted in revised form: 21 September 2000 / Published online: 21 November 2000 Ó Springer-Verlag 2000
Abstract Objectives: Assessment of predictors for initiating prophylactic antithrombotic prescribing for patients newly diagnosed with ischaemic heart disease (IHD) and the impact of the introduction of national guidelines. Design and setting: A retrospective case±control study was performed using pharmacy prescription data from 120,000 Dutch patients over a 5-year period. IHD patients were identi®ed using as a marker multiple nitrate prescriptions [anatomical-chemical-therapeutic (ATC)
R. Fijn (&) á A. G. G. Stuurman-Bieze á P. B. van den Berg J. R. B. J. Brouwers á L. T. W. de Jong-van den Berg Groningen University Institute for Drug Exploration (GUIDE), University of Groningen, University Centre for Pharmacy, Department of Pharmaceutical Pharmacology and Clinical Pharmacy, Section of Social Pharmacy and Pharmacoepidemiology, Groningen, The Netherlands Tel.: +31-50363-3261; Fax: +31-50363-2772 e-mail:
[email protected] A. G. G. Stuurman-Bieze Wittesteijn Pharmacy, Emmeloord, The Netherlands A. G. G. Stuurman-Bieze á P. B. van den Berg L. T. W. de Jong-van den Berg InterAction working group, Northern Netherlands, The Netherlands J. R. B. J. Brouwers General Hospital De Tjongerschans, Department of Clinical Pharmacology and Hospital Pharmacy, Heerenveen, The Netherlands P. A. de Grae University Hospital Groningen and University of Groningen, Faculty of Medical Sciences, Department of Clinical Pharmacology, Groningen, The Netherlands R. Fijn Medical Sciences and Pharmacy, Social Pharmacy and Pharmacoepidemiology, Antonius Deusinglaan 1, 9713 AV Groningen, The Netherlands
code C01D] indicating chronic use. Initiation of antithrombotic therapy was likewise identi®ed using ATC codes B01AA and B01AC (oral anticoagulants and thrombocyte aggregation inhibitors), prescribed within 6 months following the ®rst nitrate prescription. Statistically signi®cant (P1, anatomical-chemical-therapeutic (ATC) code C01D] indicating chronic use. The type of nitrate prescribed ®rst as maintenance therapy (versus rescue therapy) was considered to be the main nitrate therapy. To select newly diagnosed patients (incident patients), a nitrate-free period of 6 months preceding the ®rst nitrate prescription was de®ned. Also a possible follow-up period of 6 months following the ®rst nitrate prescription was de®ned. Therefore, the actual research period comprised 1 July 1994 until 30 June 1998. Consequently, 2598 patients met the inclusion criteria and were selected as newly diagnosed IHD patients. Validation of this identi®cation procedure for IHD patients (MI and/or AP) using (Dutch) pharmacy prescription data has shown to be over 90% speci®c and sensitive, nonetheless still including patients with only heart failure [2, 6, 44, 45]. For the 2598 patients, prophylactic antithrombotic therapy was identi®ed using as a marker multiple prescriptions (>1) for low-dose thrombocyte aggregation inhibitors (no more than 300 mg acetylsalicylic acid and carbasalate calcium) and multiple prescriptions (>1) for oral anticoagulants (acenocoumarol and phenprocoumon), with ATC codes B01AC and B01AA, respectively. Using multiple (>1) prescriptions for other pharmacotherapy identi®ed possible cardiovascular (other than IHD), respiratory, diabetic and gastrointestinal co-morbidity. A retrospective case±control study was performed. In terms of outcome, patients were considered cases if antithrombotic therapy was initiated within 6 months following the ®rst nitrate prescription. Patients were considered controls if antithrombotic therapy was not initiated within 6 months. Data were univariably and multivariably (stepwise backward logistic regression) analysed. Crude odds ratios (ORcrude) with 95% con®dence intervals (95% CI) and P-values (v2 testing) were calculated to determine whether patient gender and age, GP, the number of drugs used simultaneously, possible co-morbidity, main type of nitrate therapy and period of diagnosis (in particular whether the patient was diagnosed before or after the national guidelines' introduction) would be predictors for prescribing antithrombotics. Possible predictors were categorised where relevant. Statistical signi®cance was set at 0.05. All data were processed and analysed utilising MSAccess 7.0 and SPSS 9.0.
Results Table 1 shows the cross-sectional (5-year period) baseline characteristics of all eligible subjects and of newly
741 Table 1 Baseline characteristics of eligible subjects. All ischaemic heart disease (IHD) patients including newly diagnosed (incident) and prevalent patients and the subselection of newly diagnosed IHD patients only. ISDN isosorbid dinitrate; ISMN isosorbid mononitrate; NG nitroglycerine
Characteristics
Gender Male Female Agea (years, mean 68.2, range 27±107) 90 Main type of nitrate therapyb ISDN ISMN NGc Antithrombotic therapy None Thrombocyte aggregation inhibitors (acetylsalicylic acid, carbasalate calcium) Oral anticoagulants (acenocoumarol, phenprocoumon) Thrombocyte aggregation inhibitors and oral anticoagulants Period of diagnosis Second half 1994 First half 1995 Second half 1995 First half 1996 Second half 1996 First half 1997 Second half 1997 First half 1998 Guidelines introduction (®rst half of 1996) Pre-guidelines period Post-guidelines period Co-medication Cardiovascular pharmacotherapyd Pulmonary pharmacotherapy Antidiabetics Antacids/antiulcerics
All (newly diagnosed and prevalent, n=5187) n (%)
Newly diagnosed (n=2598) n (%)
2850 (54.9) 2337 (45.1)
1401 (53.9) 1197 (46.1)
62 336 819 1340 1732 813 85
(1.2) (6.5) (15.8) (25.8) (33.4) (15.7) (1.6)
37 204 460 693 825 346 33
(1.4) (7.9) (17.7) (26.7) (31.8) (13.3) (1.3)
2662 (51.3) 850 (16.4) 1675 (32.3)
1256 (48.3) 426 (16.4) 916 (35.5)
1878 (36.2) 2508 (48.4)
913 (35.1) 1326 (51.0)
616 (11.9)
255 (9.8)
185 (3.6)
104 (4.0) 343 352 320 328 307 290 313 345
(13.2) (13.5) (12.3) (12.6) (11.8) (11.2) (12.0) (13.3)
1255 (48.3) 1343 (51.7) 4339 655 589 1182
(83.7) (12.6) (11.4) (22.8)
2147 348 300 634
(82.6) (13.4) (11.5) (24.4)
a
Age at the date of diagnosis (i.e. index date of ®rst nitrate prescription) Registered for ischaemic heart disease in the Netherlands; oral (ISMN, ISDN, NG) or transcutaneous (ISDN, NG) maintenance therapy, eventually supplemented with oromucosal (ISDN, NG) rescue therapy c Oral or transcutaneous maintenance therapy, or oromucosal rescue therapy combined with oral b-blockers or calcium channel blocking agents maintenance therapy d Other than nitrate and antithrombotic therapy b
diagnosed IHD patients only. Analysis of gender and age distribution showed that male patients aged 70± 79 years were most represented. During the study period, isosorbide dinitrate (ISDN) as the main type of nitrate therapy decreased (55.1% in 1994, 47.5% in 1998). Isosorbide mononitrate (ISMN) as the main type of nitrate therapy increased (11.9% in 1994, 20.9% in 1998). The percentage of nitroglycerine (NG) remained stable (33.0% in 1994, 31.6% in 1998). Nonetheless, ISDN was favoured overall as the main nitrate therapy. Further analysis showed that about 35% of all patients were not prescribed any type of antithrombotics. Exploration of co-prescribing with the view to derive possible co-morbidity showed that a high proportion used cardiovascular medications other than nitrates and
antithrombotics. With respect to the type of antithrombotic therapy, the ratio of thrombocyte aggregation inhibitors to oral anticoagulants remained stable at 5:1 (78%:16%). Although elderly patients were less likely to use oral anticoagulants instead of thrombocyte aggregation inhibitors, there were no statistically signi®cant ®ndings in the distribution of the type of antithrombotic therapy across subcategories of characteristics. Generally, characteristics of newly diagnosed IHD patients were similar to baseline characteristics of all eligible subjects. The number of patients newly diagnosed in the post-guidelines period (n=1343, 51.7%) was slightly higher than the number of patients diagnosed in the pre-guideline period (n=1255, 48.3%).
742
However, distribution of patients across the eight subcategories of diagnosis period was approximately equal (range 290±352, mean 325 per period). Although the distribution of patients across all GPs is not displayed, dierences in antithrombotic prescribing for their newly diagnosed patients were substantial (range 29.9±80.9%). Figure 1 visualises antithrombotic prescribing over time including the time of guidelines introduction. Table 2 shows the results of the univariable and multivariable analysis, considering all possible predictors. Possible pulmonary and gastrointestinal co-morbidity and the period of diagnosis were not statistically signi®cant predictors after univariable analysis. Also, after multivariable analysis, patient age, GP and the number of drugs used simultaneously were not statistically signi®cant predictors. Gender, possible cardiovascular (other than IHD) and diabetic co-morbidity, the main type of nitrate therapy and whether diagnosis was set before or after the introduction of national guidelines were all statistically signi®cant predictors. GPs were 2.4 times more likely to prescribe antithrombotics for male than for female patients. Prescribing antithrombotics was about 6.4 and 1.6 times more likely for patients with cardiovascular diseases other than IHD and diabetes mellitus, respectively. Patients using ISMN as the main therapy were 2.1 times more likely to be prescribed antithrombotics than patients using ISDN. Antithrombotic prescribing for patients using NG as the main therapy was 1.3 times less likely than patients using ISDN. Finally, the introduction of national guidelines increased the chance of antithrombotic prescribing by a factor of 1.4.
Discussion The present ®ndings indicate that for about 65% of all newly diagnosed IHD patients, antithrombotics are initiated within 6 months after diagnosis. Regarding the type of antithrombotic therapy, thrombocyte aggregation inhibitors are favoured over oral anticoagulants. Male patients and patients using ISMN as the main type of nitrate therapy are more likely to be prescribed antithrombotics than female patients and patients using ISDN or NG as the main type of nitrate therapy. Car-
Fig. 1 Trends in antithrombotic prescribing in ischaemic heart disease (IHD) patients (1994±1998)
diovascular (other than IHD) and diabetic co-morbidity are additional predictors for initiating antithrombotic therapy. Finally, the introduction of national guidelines has resulted in an increase in antithrombotic prescribing. Although identi®cation of morbidity based on pharmacy prescription data is known for its varying sensitivity and speci®city depending on the drug groups used as markers, identi®cation of IHD using multiple nitrate prescriptions has shown to be both highly sensitive and speci®c [2, 6, 44, 45]. Another strength is the fact that low-dose thrombocyte aggregation inhibitors are `prescription only' and not `over-the-counter', contrary to abroad. Nonetheless, a proportion of patients who have been included may have used nitrates for indications other than IHD, such as heart failure or achalasia. This did not interfere with the research design in the case of patients suering from both IHD and heart failure. Others estimated heart failure among nitrate users at approximately 10%, roughly identi®ed by digoxin use. Exclusion from this study of these nitrate users with only heart failure might have slightly increased the proportion of IHD patients being prescribed antithrombotics. Furthermore, NG may be used as a diagnostic agent as `heartburn' and acute IHD are symptomatically alike. Using multiple prescriptions for nitrates circumvents much of this possible misclassi®cation. However, a slight proportion of patients with IHD that have not been diagnosed as such, or that have been diagnosed as such but that may not have been prescribed nitrates in view of adverse eects, will have been missed. These patients frequently use b-blockers or calcium channel blocking agents instead of long-acting nitrates as anti-ischaemic maintenance therapy, most often supplemented with a sublingual nitrate for rescue therapy and have therefore still been included. The potential misclassi®cation also applies to identi®cation of antithrombotic therapy. Although some patients may have used oral anticoagulants for indications other than IHD prophylaxis, such as thrombosis in general, this serves as IHD prophylaxis as well and as such does not interfere with the research design. Some of the IHD patients may have used thrombocyte aggregation inhibitors for indications other than IHD prophylaxis, such as transient ischaemic attack (TIA) and cerebrovascular accident (CVA) prophylaxis, claudicatio intermittens or for antiphlogistic and analgesic purposes. Again, in the former two cases, this indirectly serves as IHD prophylaxis as well and as such does not interfere with the research design. In the latter case it concerns prescribing high-dose thrombocyte aggregation inhibitors (more than 300 mg) that are characterised by a dierent ATC code (N02BA). The use of high-dose thrombocyte aggregation inhibitors may have been a consideration for not initiating additional antithrombotics. This was checked but none of the included patients not being prescribed antithrombotics used chronic high-dose thrombocyte aggregation inhibitors, indicating no interference with the research design. The increase over time in prescribing antithrombotics is similar to that seen abroad [43]. Still, the present
743 Table 2 Predictors for the initiation of antithrombotic prescribing in newly diagnosed ischaemic heart disease (IHD) patients (n=2598). NS not signi®cant; ISDN isosorbid dinitrate; ISMN isosorbid mononitrate; NG nitroglycerine; CI con®dence interval; OR odds ratio Explanatory variables (predictor)
Gender Male Female Age (years) Overall 90 General practitioner (n=59) Overall No. simultaneously used drugs Overall 8 Co-morbiditya Cardiovascular Pulmonary Diabetes Gastrointestinal Main type of nitrate therapy Overall ISDN ISMN NG Period of diagnosis Overall Second half 1994 First half 1995 Second half 1995 First half 1996 Second half 1996 First half 1997 Second half 1997 First half 1998 Introduction of guidelines Pre-guidelines period Post-guidelines period
Univariable analysis
Multivariable analysis
ORcrude
95% CI
P value
ORadjusted
95% CI
P value
1.0 0.480
± 0.408±0.565
