PHARMO Record Linkage System Research Articles

Haemoglobin A1c Goal Attainment in Relation to Dose in Patients with Diabetes Mellitus Taking Metformin

Conclusions from clinical trials suggest possible therapeutic advantages for once-daily agents over twice-daily agents in the treatment of type 2 diabetes mellitus. This study set out to investigate the relationship between metformin dosing frequency and glycosylated haemoglobin (HbA1c)-goal attainment in daily practice. This was a nested case-control study. Data were obtained from the PHARMO Record Linkage System, which includes linked drug-dispensing and clinical-laboratory records for approximately three million individuals in defined areas of the Netherlands. The study cohort included new users of oral antihyperglycaemic drugs between 1999 and 2005 with a baseline HbA1c> or =7% and at least one further HbA1c measurement within 18 months after the index date. Cases attained HbA1c goal (<7%) within 18 months; controls did not attain this HbA1c goal. Compliant cases and controls taking metformin monotherapy were included in the analyses. Dosing frequency was dichotomized into once daily and twice daily or more frequently. In the multivariate analysis we considered oral antihyperglycaemic dose, baseline HbA1c, first prescriber and number of HbA1c measurements. The study cohort included 3107 new oral antihyperglycaemic drug users. The analyses included 753 cases and 477 controls taking metformin. Dosing twice daily or more was associated with a 71% higher probability of attaining goal (odds ratio 1.71 [95% CI 1.31, 2.24]) compared with once-daily dosing, after adjustment for baseline HbA1c, first prescriber, sex and age. We could not distinguish between the effect of dose and dosing frequency as these were closely related. Statistical testing in the analyses stratified by dose was prohibited by small numbers. About 40% of compliant metformin users did not achieve their HbA1c goal within 18 months because of dosing problems. However, the strong correlation between total daily dose and dosing frequency did not permit identification of which of these dosing issues was the most important contributor to not achieving HbA1c goal.

Use of inhaled corticosteroids and the risk of non-fatal acute myocardial infarction

Use of inhaled corticosteroids may reduce the risk of acute myocardial infarction (MI) through reductions in systemic inflammation and C-reactive protein. To examine the association between the use of inhaled corticosteroids and the risk of non-fatal acute MI. In the Dutch PHARMO record linkage system database, we conducted a case-control study (2476 MI cases), nested in a cohort of antihypertensive drug users. The use of inhaled corticosteroids 100 days before the index date was compared with never use. We adjusted the analyses for the severity of the underlying respiratory disease and general drug and disease history. We found that the use of inhaled corticosteroids was not associated with a decreased risk of non-fatal MI in antihypertensive drug users after adjustment for the underlying respiratory disease severity, adjusted odds ratio (OR) 1.24, 95% confidence interval (CI) 0.97-1.57. A higher daily dose (adjusted OR 1.82, 95% CI 0.80-4.13) and longer duration of use (adjusted OR 1.28, 95% CI 0.90-1.81) were not associated with a decreased risk of non-fatal MI. An inhaled corticosteroid dispensing in the 30 days before the index date was not protective but resulted in a 1.7-fold increased risk of non-fatal MI. Our results do not support the hypothesis that inhaled corticosteroids protect against the risk of non-fatal MI by a reduction of systemic inflammation.

The effect of discontinuation of antihypertensives on the risk of acute myocardial infarction and stroke

ABSTRACTObjective: Adherence to antihypertensive drug treatment is suboptimal. The present study investigates the effect of early treatment discontinuation with antihypertensive drugs on the risk of acute myocardial infarction (AMI) or stroke in daily clinical practice.Methods: In the PHARMO Record Linkage System, which includes all records of drug dispensings and hospitalisations for ≥2 million subjects in the Netherlands, new users of antihypertensive (AHT) drugs ≥18 years of age were studied during the period 1 January 1993 – 1 October 2002 to determine the risk of AMI or stroke related to persistence with AHT. Patients were initially followed for 2 years to determine persistence with AHT, and then for a further 2 years or until the first hospital admission for AMI or stroke, death, or end of the study period. Patients using AHT for secondary prevention of cardiovascular disease were excluded from the study cohort.Results: The study cohort included 77 193 AHT users. The percentage of non-persistent patients was 55% at 2 years, with the lowest non-persistence rates for angiotensin-receptor blockers (ARBs) and ACE-inhibitors (40%) and the highest rates for β-blockers, calcium-channel blockers (CCBs) and diuretics (54–61%). Non-persistent AHT use was associated with a 15% increased risk of AMI (RR 1.15; 95% CI 1.00–1.33) and a 28% increased risk of stroke (RR 1.28; 95% CI 1.15–1.45).Conclusions: The results of this study show that in daily clinical practice early discontinuation of antihypertensive drug treatment in primary prevention increases the risk of subsequent AMI or stroke.

PDB69 DETERMINANTS OF STOPPING TREATMENT WITH ORAL ANTIDIABETIC DRUGS IN DAILY CLINICAL PRACTICE

DETERMINANTS OF STOPPINGTREATMENTWITH ORAL ANTIDIABETIC DRUGS IN DAILY CLINICAL PRACTICE Koerselman J, van der Bij S, Erkens JA, Kessabi S, Groot MT, Penning-van Beest FJA, Herings RMC PHARMO Institute, Utrecht, Utrecht,The Netherlands, Novartis Pharma AG, Basel, Switzerland, Novartis Pharma B.V, Arnhem,The Netherlands OBJECTIVES: In daily practice many Type 2 Diabetes Mellitus (T2DM-)patients discontinue treatment. Therefore, the aim of this study was to investigate the determinants of (non-) persistence with oral antidiabetic drugs (OADs) in daily clinical practice. METHODS: From the PHARMO record linkage system, comprising among others linked drug-dispensing, and hospital data for >2.3 million subjects in The Netherlands, new users of OADs were identified in the period 1999–2005. Patients with 1 year of follow-up, were included in the study-cohort. Persistence with OAD-treatment in the first year of therapy was determined using the method of Catalan. Potential determinants of (non-)persistence included patient-characteristics, type of initial OAD-therapy, and cardiovascular co-morbidity. RESULTS: The study included 33,299 new users of OADs. One year after start, 42% of new T2DM-patients had stopped using any OAD. The risk of non-persistence was decreased with male gender (HR: 0.97; 95% CI: 0.94–1.00), and cardiovascular drug use (HR: 0.91; 95% CI: 0.86–0.97). Regarding age, compared to patients 76 years, the age-group 55–75 years had a 16% lower risk, and the age-group 30–44 years had a 32% increased risk of non-persistence. Patients starting on combined metformin + SU had a lower risk of non-persistence with any OAD; compared to patients starting on metformin monotherapy, the risk was 23% lower (HR: 0.77; 95% CI: 0.70–0.85). The risk of nonpersistence was increased with a specialist as first prescriber (HR: 1.20; 95% CI: 1.15–1.26), higher initial daily dose (HR: 1.09; 95% CI: 1.00–1.22), and higher initial daily dosing frequency (HR: 1.10; 95% CI: 1.02–1.20). CONCLUSION: In daily clinical practice about 40% of new T2DM-patients stop OADtherapy within one year. Determinants of stopping OADmedication were male gender, age-group, specialist as first prescriber, dosing, cardiovascular drug-use, and type of initial OAD-treatment.

Society already achieves economic benefits from generic substitution but fails to do the same for therapeutic substitution

To assess the potential annual savings due to generic and therapeutic substitution of statin therapy for the general Dutch population, taking the patients medical history into account. We conducted a population-based costing study using the PHARMO Record Linkage System (RLS). PHARMO RLS contains drug dispensing records from a representative sample of pharmacies located in more than 50 regions in the Netherlands. We selected all statin users in the database since 2003. The cost-savings of generic substitution of statin therapy for all simvastatin and pravastatin users, and of therapeutic substitution of statin therapy for other statin users were calculated. Substituting current users and new users of statins were considered separately. Therapeutic substitution was based on the medical history of the individual patient. Patients were only substituted if there was an appropriate substitute available. The appropriateness of substitution was based on drug-drug interactions between statins and possible comedication and the availability of an equipotent alternative. Substituting (generic and therapeutic) statin therapy for all current users would lead to potential annual savings of approximately 87 million euros. Substituting (generic and therapeutic) all starters on statin therapy would lead to potential annual savings of around 51 million euros. In the case of generic substitution only, the potential annual savings for all current simvastatin and pravastatin users would be 2.4 million euros and for the new users about 1.8 million euros. From an economic point of view, society could gain a lot from substituting statin therapy, especially from therapeutic substitution.

Quantity and quality of potential drug interactions with coumarin anticoagulants in the Netherlands

Coumarin anticoagulants are prone to potentially life-threatening drug-drug interactions due to a combination of unfavorable properties. However, real life data on the actual occurrence are scarce. The aim of this study was to quantify and qualify potential drug interactions with coumarin anticoagulants in daily practice. A cohort study including all users of phenprocoumon or acenocoumarol during the period 1991-2003 in the PHARMO Record Linkage System. All 24 individual drugs and 11 drug groups interacting with coumarins according to central database used in the Dutch pharmacies were considered. Frequency and type of potential drug interactions during anticoagulant therapy with coumarins. 48,627 out of 76,455 mainly acenocoumarol-users (64%) were dispensed at least one potentially interacting drug (PID) during anticoagulant therapy. About 35% of these cases were dispensed a (very) strongly interacting drug, whereas 3% were dispensed a contraindicated drug. Antibacterial drugs and NSAIDs (39% and 37% of all users, respectively) were the most frequently dispensed PIDs. Potential drug interactions with coumarins frequently occur in daily practice, confronting two-thirds of patients with an increased risk of bleeding. To a large part, this is attributable to commonly prescribed medication like antibacterial drugs and NSAIDs. This situation substantiates the need for proper monitoring or new anticoagulants with less drug-drug interactions.

Use of β-Blockers and the Risk of Hip/Femur Fracture in the United Kingdom and The Netherlands

Data from in vivo studies have indicated a role for β-blockers in the prevention of bone loss. Some epidemiological studies have found protective effects of β-blockers on fracture risk. However, there is limited information on the association with cumulative dose and type of β-blockers used. We conducted two case-control studies using data from the UK General Practice Research Database (GPRD) and the Dutch PHARMO Record Linkage System (RLS). Cases were patients with a first hip or femur fracture; controls were individually matched on practice/region, gender, year of birth, and calendar time. Current use of β-blockers was defined as a prescription in 90 days before the index date. We adjusted for medical conditions and drugs associated with falling or bone mineral density. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using conditional logistic regression analysis. The study population included 22,247 cases and controls in the GPRD and 6,763 cases and 26,341 controls in the PHARMO RLS. Current use of β-blockers was associated with a reduced risk of hip/femur fracture in both the GPRD (adjusted OR = 0.82, 95% CI 0.74–0.91) and PHARMO RLS (adjusted OR = 0.87, 95% CI 0.80–0.95) study populations. However, this reduction of risk was not associated with cumulative dose, lipophilicity, or receptor selectivity of β-blockers. The protective effect of β-blockers was only present among patients with a history of use of other antihypertensive agents (GPRD adjusted OR = 0.72, 95% CI 0.64–0.83; PHARMO RLS adjusted OR = 0.76, 95% CI 0.67–0.86) but not in patients using β-blockers only (GPRD adjusted OR = 0.97, 95% CI 0.82–1.14; PHARMO RLS adjusted OR = 1.01, 95% CI 0.90–1.14). Also, in patients with a history of use of other antihypertensive agents, no dose-response relationship with β-blocker use was found. The effect was constant with cumulative dose and the OR was below 1.0 even among patients who just started treatment with β-blockers. As the mechanism by which β-blockers could influence bone mineral density is likely to need some time to exert a clinically relevant effect, all these finding suggests that the association between β-blockers and fracture risk is not causal.

Influence of initial use of serotonergic antidepressants on antiparkinsonian drug use in levodopa-using patients

To assess whether there is an association between initial use of serotonergic antidepressants and changes in antiparkinsonian drug treatment. A retrospective cohort study was performed with the PHARMO record linkage system. All patients from 1994 until 2004 of 40 years or older who were first time users of an antidepressant and who had used a levodopa-containing drug at least 180 days before initiation of the antidepressant were included. The maximum follow-up time was 180 days. The first change in antiparkinsonian drug treatment, defined as an increase in the daily dosage of any antiparkinsonian drug, the start of a new antiparkinsonian drug or a change in the dosage form during the follow-up period, was taken as an endpoint. Antidepressants were classified in two ways: according to their class [selective serotonin reuptake inhibitors (SSRI), tricyclic antidepressants (TCA) or other antidepressants] or by the extent of their inhibition of serotonin reuptake (high, intermediate or low). A total of 221 patients was included in our study. The adjusted hazard ratio for a change in antiparkinsonian drug treatment was 0.7 (95% CI 0.3-1.5) comparing SSRI with TCA users, and it was 0.9 (95% CI 0.4-2.1) comparing users of other antidepressants with TCA users. The adjusted hazard ratio for a change in antiparkinsonian drug treatment was 0.6 (95% CI 0.3-1.4) comparing users of antidepressants with high versus low extent of inhibition of serotonin reuptake, and it was 0.7 (95% CI 0.3-1.4) comparing users of antidepressants with intermediate versus low extent of inhibition of serotonin reuptake. Based on these observations, we found no evidence to be more cautious using SSRIs or serotonergic antidepressants compared to other antidepressants in patients with Parkinson's disease.

Changes in medication associated with epilepsy‐related hospitalisation: a case‐crossover study

To assess the association between changes in medication and epilepsy-related hospitalisation. Data were obtained from the PHARMO Record Linkage System (Jan 1998 to Dec 2002). We conducted a case-crossover study among patients with a first epilepsy-related hospital admission who had continuously used at least one antiepileptic drug (AED) during a 28-week period before admission. For each patient, changes in medication in a 28-day window before hospitalisation were compared with changes in four earlier 28-day windows. Evaluated changes were: changes in AEDs (pattern and dosage), changes in interacting co-medication and changes in non-interacting co-medication (i.e. introduction of non-interacting drugs). The strength of the association between changes in medication and epilepsy-related hospitalisation was estimated using conditional logistic regression analysis and expressed as odds ratios (ORs) with 95% confidence intervals (CI). Out of 1185 patients with a first epilepsy-related hospitalisation, 217 patients met the inclusion criteria. Of the changes in antiepileptic therapy, discontinuation showed a trend towards an increased risk of hospitalisation (OR: 2.57; 95%CI: 0.81-8.17). Drug interactions influencing antiepileptic therapy rarely occurred. Introduction of three or more non-interacting drugs was significantly associated with epilepsy-related hospitalisation (OR: 4.80; 95%CI: 2.12-10.87). Of individual drugs, addition of antimicrobial agents was significantly associated with epilepsy-related hospitalisation (OR: 1.99; 95%CI: 1.06-3.75). Changes in AED therapy were not significantly associated with epilepsy-related hospitalisation and few drug interactions influencing antiepileptic therapy occurred. However, patients starting three or more new non-AEDs had a nearly five times increased risk of epilepsy-related hospital admission.

Risk of ischemic complications related to the intensity of triptan and ergotamine use

To investigate whether the intensity of triptan and ergotamine use, in specific overuse, is associated with the risk of ischemic complications. We conducted a retrospective nested case-control study using data from the PHARMO Record Linkage System. All patients with more than one prescription for either a triptan or ergotamine were initially identified. Cases were all patients who were admitted to the hospital for an ischemic complication. Matched controls were assigned the same index date as the cases. The determinant was the intensity of use of triptans and ergotamine during 1 year preceding the index date. Overuse was defined as use of > or =90 defined daily doses during that year. Conditional logistic regression was used to estimate odds ratios (ORs), adjusting for confounders. Stratified analysis was used to estimate the risk for both patients using and those not using cardiovascular drugs. A total of 17,439 patients received more than one prescription. A total of 188 cases and 689 controls were identified. Triptan overuse was not associated with an increased risk of ischemic complications (OR 0.96; 95% CI: 0.49 to 1.90). Overuse of triptans in patients concomitantly using cardiovascular drugs did not increase this risk. Overuse of ergotamine turned out to be a risk factor for ischemic complications (OR 2.55; 95% CI: 1.22 to 5.36). Patients overusing ergotamine and concomitantly using cardiovascular drugs were at highest risk (OR 8.52; 95% CI 2.57 to 28.2). In general practice, triptan overuse does not increase the risk of ischemic complications. Overuse of ergotamine may increase the risk of these complications, especially in those simultaneously using cardiovascular drugs.

Persistent bisphosphonate use and the risk of osteoporotic fractures in clinical practice: a database analysis study

ABSTRACTObjective: This study aimed to investigate the association between persistent use of bisphosphonates and the risk of osteoporotic fractures in clinical practice.Methods: Data were obtained from the PHARMO Record Linkage System, which includes, among other databases, drug-dispensing records from community pharmacies linked to hospital discharge records of more than two million subjects in defined areas in the Netherlands. Persistence with bisphosphonate therapy was assessed during a period of 3 years. A nested matched case control study (cases:controls = 1:10) was performed to study the association between persistent bisphosphonate use and hospitalisation for osteoporotic fractures and analysed by conditional logistic regression analysis. The analyses were adjusted for patient characteristics such as previous hospitalisations for fractures, co-morbidity and co-medication.Results: 14 760 new female users of bisphosphonates were identified of which 541 women had a hospitalisation for osteoporotic fracture after start of bisphosphonate treatment (1–3 years follow-up). One-year persistence rates increased from 33% with alendronate daily to 48% with alendronate weekly, an increase of 15%. Similar results were obtained with risedronate daily and weekly. One year persistent use of bisphosphonates resulted in a statistical significant 26% lower fracture rate (OR 0.74; 95%CI 0.57–0.95) whereas 2 year persistent use resulted in a 32% lower rate (OR 0.68; 95%CI 0.47–0.96).Conclusions: Persistent use of bisphosphonates decreases the risk of osteoporotic fractures in clinical practice. Approximately 6% of fractures among users of bisphosphonates could be prevented if persistence was improved by 20%. However, current persistence with bisphosphonate therapy is suboptimal and strategies that further increase persistence are likely to further prevent the number of fractures.

Angiotensin-converting enzyme inhibitor use and pneumonia risk in a general population

The aim of the present study was to assess whether the use of angiotensin-converting enzyme (ACE) inhibitors is associated with a decreased risk of hospitalisation for community-acquired pneumonia (CAP) in a general, essentially white population. Data were obtained from the Dutch PHARMO Record Linkage System. Cases were defined as patients with a first hospital admission for CAP. For each case, up to four population controls were matched by age and sex. The study population comprised 1,108 patients with a first hospital admission for CAP and 3,817 matched controls. After adjusting for several confounders, ACE inhibitor use was not associated with a decreased incidence of pneumonia (adjusted odds ratio (OR) 1.12; 95% confidence interval (CI) 0.88-1.43). Additionally, no significant association was observed in patients with diabetes, respiratory diseases, heart failure, or patients with both of the last two conditions. Furthermore, adjustment of treatment effects on pneumonia risk using stratification on balancing score also showed no significant association between ACE inhibitor use and pneumonia risk within the different strata (overall adjusted OR 1.09; 95% CI 0.87-1.36). In contrast with previous findings in Asian populations, the current authors were not able to confirm the beneficial effect of angiotensin-converting enzyme inhibitors on pneumonia risk in a general, essentially white population.

Improving Pharmacotherapy After Myocardial Infarction by Group Academic Detailing Using Feedback Data on a Patient Level

To develop and evaluate a peer review group (PRG) meeting using feedback data on a patient level to improve the quality of drug therapy for prevention of recurrent myocardial infarction. Prospective follow-up study. General practitioners' computerized patients records (intervention patients) and the PHARMO record linkage system (controls). Forty patients in the intervention group and 1030 control patients; both groups had documented myocardial infarction. The intervention, which was based on the principles of group academic detailing, consisted of scoring current cardiovascular treatment on separate forms for each patient, presenting an overview of, and discussing, evidence-based treatment after myocardial infarction, defining the target population, formulating a binding consensus, and identifying patients who were eligible for improvement of pharmacotherapy. Drug therapy and adherence to the newly formulated PRG consensus were assessed at baseline and 1 year after the intervention. Of the patients who received the intervention and were not treated according to the PRG consensus at baseline, 40% received treatment according to the consensus 12 months after the PRG meeting. In the control group, the proportion of patients was 9.5% (prevalence ratio 4.2, 95% confidence interval 1.8-9.7). Peer review group meetings can be a valuable tool for improving pharmacotherapy after myocardial infarction.

Determinants of persistence with bisphosphonates: A study in women with postmenopausal osteoporosis

Background: Although bisphosphonates are useful in the management of osteoporosis, patients often discontinue treatment. Objectives The aims of this study were to investigate persistence with bisphosphonates, and to assess whether the dose interval influenced persistence, among women with postmenopausal osteoporosis. Methods Data were obtained from the PHARMO Record Linkage System, which includes, among other databases, drug-dispensing records from community pharmacies linked to hospital discharge records of >1 million subjects in defined areas in The Netherlands. Women who were new users of alendronate (daily or weekly), etidronate (daily), or risedronate (daily) during the period from January 2000 through September 2003 were eligible for inclusion in the study if they were aged ≥55 years or had been hospitalized for a menopausal disorder. One-year rates of persistence with treatment (defined as the percentage of patients who used the drug for ≥365 days without failure to continue renewals) were determined by using episodes of bisphosphonate treatment. The association between persistence and dose intervals, type of bisphosphonate, and other determinants (including age, occurrence of gastrointestinal adverse events as measured by use of concomitant medications [eg, antacids, proton pump inhibitors, histamine 2 (H 2)-receptor antagonists, misoprostol, laxatives, antidiarrheals, bowel motility enhancers] and fractures) was assessed. To study whether persistence with bisphosphonates was associated with the former use of other antiosteoporosis medication or the presence of drug-induced osteoporosis, the use of hormone replacement therapy, raloxifene, and systemic corticosteroids in the 6 months before the index date were included as determinants. Results The study sample included 2124 women who were new users of bisphosphonates. The mean (SD) age of the study population was 71.6 (8.7) years. After 1 year, 51.9% of weekly alendronate users and 30.1% to 42.2% of daily bisphosphonate users were persistent. In the multivariate analysis (which included age, concomitant medication, and fractures), patients using alendronate weekly were significantly more likely to persist than those using alendronate daily (relative risk [RR], 1.56 [95% CI, 1.32–1.85]). The likelihood of persistence was similar among those who used the daily regimens of risedronate, etidronate, and alendronate. The occurrence of gastrointestinal adverse events was associated with decreased persistence with bisphosphonates (Hin2)-receptor antagonists: RR, 0.71 [95% CI, 0.53–0.94]; bowel motility enhancers: RR, 0.78 [95% CI, 0.65–0.94]). Conclusions In this study, dose interval and the occurrence of gastrointestinal adverse events were independent determinants of persistence with bisphosphonate therapy. Although the likelihood of persistence with bisphosphonate use was significantly higher among those who used a less frequently administered regimen, persistence rates were still suboptimal.

Regional Influences on the Dispensing of Glucose Test Strips in Dutch Community Pharmacies

Pharmacy practice guidelines promote the role of community pharmacies in self-monitoring of blood glucose. However, variations between Dutch pharmacies exist in the proportion of patients to whom test strips are dispensed. To assess whether variations between community pharmacies in dispensing of blood glucose test strips can be explained by differences in patient characteristics and the region in which the pharmacy is located. PHARMO-Record Linkage System containing drug dispensing histories from 40 community pharmacies of about 450,000 patients in the Netherlands. We performed a retrospective cohort study. Data on prescription of all drugs and medical aids between 1991 and 2001 were extracted for all new users of antidiabetic drugs. Patients were excluded if the dispensing history did not allow classification of the type of diabetes or if the dispensing pharmacy could not be determined. The data were analysed using a Cox proportional hazard model. Time to first test strips dispensed. We identified 8,233 starters of antidiabetic drugs. During a median follow-up of 2.1 years, 20% of the patients were dispensed test strips at least once. Community pharmacy was significantly associated with the dispensing of test strips after adjustment for patient characteristics. This association was less apparent when stratified for the geographical location of the pharmacy. Community pharmacy is an independent determinant of the start of use of test strips. Differences in dispensing of test strips between pharmacies are dependent on geographical region. This suggests that implementing practice guidelines for diabetes care in community pharmacy requires different approaches in different regions.

Drug-Induced Thrombocytopenia

Drug-induced immune thrombocytopenia, excluding heparin-induced thrombocytopenia, is a rare adverse drug reaction for which the evidence about frequency, relative risk and risk factors mainly originates from case reports and case studies. This study aims to quantify the risk for thrombocytopenia following exposure to drugs that are most often reported to cause thrombocytopenia in the general population. A retrospective, case-control study was conducted within the PHARMO record linkage system. Cases were defined as patients hospitalised for thrombocytopenia in the period 1 January 1990 to 31 December 2002. For each case, up to four controls were matched based on age, sex and geographical area. Exposure on the index date to anticonvulsants, beta-lactam antibacterials, cinchona alkaloids, disease modifying antirheumatic drugs (DMARDs), diuretics, NSAIDs, sulfonamide antibacterials and tuberculostatics was assessed and categorised into mutually exclusive groups of current, recent, past and non-use. The risk was quantified with multivariate conditional logistic regression analysis. The study population comprised 705 cases and 2658 controls. Current use of beta-lactam antibacterials was associated with an increased risk for thrombocytopenia (adjusted odds ratio 7.4, 95% CI 1.8, 29.6). Increased risk estimates, although not significant, were found for current exposure to DMARDs and the sulfonamide antibacterial cotrimoxazole (trimethoprim/sulfamethoxazole). No increased risk was found for anticonvulsants, cinchona alkaloids, diuretics, NSAIDs or tuberculostatics. More evidence for an increased risk for thrombocytopenia in current use of beta-lactam antibacterials in the general population was provided. The expected increase in risk could not be confirmed for the other drugs investigated, which is possibly a result of the limited statistical power. Future studies including more patients and with laboratory data should confirm our findings before drawing definite conclusions.

Refill persistence with chronic medication assessed from a pharmacy database was influenced by method of calculation

Background and Objectives In literature, different methods of calculating persistence are used. In this study, the effect of using these different methods on persistence and the association of patients characteristics and persistence are assessed. Methods The PHARMO record linkage system was used to calculate persistence with antihypertensive drugs for a cohort of 14,466 new users of antihypertensives. Three different types of methods were used to define the maximum gap allowed between two prescriptions that a patient may have to be defined as a continuous user, one based on a defined number of days (varying from 9–365 days), the second based on the duration of the last prescription (varying from 0.1–4 times the duration), the third based on a combination of both methods, whichever leads to the lowest number of days. Results Refill persistence varied between 19.7–86.4% (method 1), between 27.9–90.2% (method 2), and between 19.7–86.4% (method 3). Furthermore, patient characteristics associated with persistence differed between and within the three different methods. Conclusion The method used and the variation within a method influenced both persistence and the association between patient characteristics and persistence. Results of persistence studies are highly influenced by the researchers' method of the maximum allowed treatment gap.

Rate and determinants of 10-year persistence with antihypertensive drugs

To assess the proportion of patients starting with antihypertensive drug treatment who continued treatment for at least 10 years. A retrospective cohort study. The PHARMO record linkage system containing drug dispensing records from community pharmacies and linked hospital discharge records of approximately 950 000 subjects. Patients who started using antihypertensive drugs (two or more prescriptions) in 1992 and did not receive a prescription for any antihypertensive drug in the 365 days preceding the first prescription. Persistence with antihypertensive drugs until 10 years. Among a total of 2325 patients who started using antihypertensive drugs, 39% used continuously during the 10 years of follow-up. Approximately 22% temporarily discontinued and restarted treatment, whereas 39% of patients discontinued permanently. Older patients were more persistent than younger patients [20-39 years: odds ratio (OR) 2.08; 95% confidence interval (CI) 1.52-2.84; 40-59 years: (reference), > or = 60 years: OR 0.69; 95% CI 0.54-0.89]. More patients who started with diuretics (reference) and beta blockers (OR 1.15; 95% CI 0.87-1.52) discontinued compared with those who started with dihydropyridine calcium antagonists (OR 0.54; 95% CI 0.34-0.84), and angiotensin-converting enzyme (ACE) inhibitors (OR 0.38; 95% CI 0.27-0.55). Patients who started with combination therapy (OR 0.29; 95% CI 0.14-0.54 compared with diuretics) or patients who were initially treated by a cardiologist (OR 0.82; 95% CI 0.61-0.97) or internist (OR 0.80; 95% CI 0.62-0.98 compared with general practitioners) also showed higher persistence. Long-term persistence in daily practice is low compared with persistence observed in randomized clinical trials and should be considered in the choice of a first-line antihypertensive agent.

Preventive drug use in patients with a history of nonfatal myocardial infarction during 12-year follow-up in The Netherlands: A retrospective analysis

Background: Myocardial infarction (MI) is a common cause of death in developed countries. Long-term preventive pharmacotherapy has been shown to decrease mortality and morbidity after MI. Based on a literature search, studies of these therapies to date have estimated the use of monotherapy, whereas many patients are prescribed combination therapy. Thus, assessment of long-term combination drug use after MI is timely. Objective: The aim of this study was to assess the use of oral antithrombotics, β-blockers, angiotensin-converting enzyme (ACE) inhibitors, hydroxymethyl-glutaryl coenzyme A reductase inhibitors (“statins”), and their combinations after MI at discharge and during 12-year follow-up. Methods: This community-based, retrospective data analysis was conducted at Utrecht University, Utrecht, The Netherlands. Data from patients aged ≥18 years at hospital admission who experienced nonfatal acute MI between 1991 and 2000 and had a duration of follow-up ≥30 days were included in the analysis. Data were retrieved from the Pharmo Record Linkage System database, which links pharmacies' dispensation records to hospitals' discharge records on an individual patient level, allowing the investigator to observe individual patients' medication use over time. Primary outcome measures were the use of pre ventive medicines (oral antithrombotics, β-blockers, ACE inhibitors, and statins) at discharge, overall use, and persistence during 12-year follow-up. Results: Of 330,000 patients in the database, 4007 were included in the analysis (2828 men, 1179 women; mean [SD] age, 63.5 [12.5] years). Use at discharge and overall use of oral antithrombotics and statins increased significantly between 1991 and 2000, whereas use of β-blockers and ACE inhibitors increased mainly in patients discharged in the latter years of the follow-up period. Therapy with any combination of drugs increased strikingly from 1991 to 2000, from 47% to 90%. At 1 year after discharge, 32% of patients had discontinued their first-prescribed combination treatments. At 5 years after discharge, this rate increased to 57%, suggesting a low rate of persistence Conclusions: Based on the results of this retrospectivedata analysis, the use of MI-preventive drug treatment at and after discharge increased significantly in this population in The Netherlands during the 1990s. Combination therapy increased strikingly. However, persistence with combination therapy was low.

Hospital discharge diagnoses of ventricular arrhythmias and cardiac arrest were useful for epidemiologic research

Objective We investigated the validity of hospital discharge diagnosis regarding ventricular arrhythmias and cardiac arrest. Methods We identified patients whose record in the PHARMO record linkage system database showed a code for ventricular or unspecified cardiac arrhythmias according to codes of the International Classification of Diseases, 9th revision, clinical modification (ICD-9-CM). The validity of ICD codes for ventricular arrhythmias and cardiac arrest (427.1, 427.4, 427.41, 427.42, 427.5, 427.69) and ICD codes for unspecified cardiac arrhythmias (427.2, 427.60, 427.8, 427.89, 427.9) was ascertained through manual review of hospital clinical records. The positive predictive value (PPV) was calculated, and differences between characteristics of true and false positives were evaluated. Results The PPV of ICD codes for ventricular arrhythmias and cardiac arrest was 82% (95% confidence interval CI = 72–92). True positive results were associated with male gender ( P = .09) and younger age ( P = .05). Of the unspecified cardiac arrhythmias 10% (95% CI = 2–18) were identified as ventricular arrhythmias or cardiac arrest. Conclusion Hospitalizations for ventricular cardiac arrhythmias and cardiac arrest (coded according to ICD-9-CM as paroxysmal ventricular tachycardia, ventricular fibrillation, ventricular flutter, ventricular premature beats, or cardiac arrest) have a high PPV and are useful for selecting events in epidemiological studies on drug-induced arrhythmias.