Adherence to gastroprotection during cyclooxygenase 2 inhibitor treatment and the risk of upper gastrointestinal tract events: a population-based study.

OBJECTIVE: Guidelines recommend coprescription of gastroprotective agents (GPAs) in patients receiving cyclooxygenase 2 inhibitors (coxibs) who are at high risk of upper gastrointestinal (UGI) tract complications (i.e., patients with a previous complicated ulcer or with multiple risk factors). Suboptimal GPA adherence has been shown to diminish the gastroprotective effect during use of nonselective nonsteroidal antiinflammatory drugs, but little is known about the effect of GPA adherence during coxib treatment. We undertook this study to determine the association between GPA adherence and UGI tract events among patients receiving coxibs. METHODS: Using primary care data from 3 databases, we conducted a case-control study in a cohort of patients age ≥50 years who were newly starting treatment with coxibs and concomitantly taking GPAs. Patients who had a UGI tract event (bleeding or symptomatic ulcer) were matched to event-free controls for age, sex, database, and calendar date. Coxib treatment intervals were defined as consecutive coxib prescriptions with intervening gaps not exceeding the duration of the previous coxib prescription. Adherence to GPAs was calculated as the proportion of days of coxib treatment covered by a GPA prescription. Odds ratios (ORs) with 95% confidence intervals (95% CIs) were calculated using conditional logistic regression analysis. RESULTS: The coxib plus GPA-treated cohort consisted of 14,416 coxib-treated patients who received GPAs for at least 1 day, yielding 16,442 coxib treatment intervals in which a GPA was coprescribed. Most patients were treated with coxibs for <30 days. Seventy-four patients had a UGI tract event during or shortly after a coxib treatment interval in which a GPA was coprescribed, with an incidence rate of 11.9 (95% CI 9.4-14.8) per 1,000 years of coxib treatment. The risk of UGI tract events was 1.97 (95% CI 0.84-4.60) for patients with <20% adherence to GPAs compared to patients with >80% adherence to GPAs. For every 10% decrease in GPA adherence, the risk of UGI tract events increased by 9% (OR 1.09 [95% CI 1.00-1.18]). CONCLUSION: Decreasing GPA adherence among coxib-treated patients is associated with an increased risk of UGI tract events.

The risk of new onset heart failure associated with dopamine agonist use in Parkinson's disease.

The aim of present study was to investigate the risk of heart failure associated with dopamine agonist use in patients with Parkinson's disease. The data sources of this study were four different population-based, healthcare databases in United Kingdom, Italy and Netherlands. A case control study nested within a cohort of Parkinson's disease patients who were new users of either dopamine agonist or levodopa was conducted. Incident cases of heart failure were identified and validated, using Framingham criteria. Controls were matched to cases on age, gender and database. To estimate the risk of newly diagnosed heart failure with ergot and non-ergot derived dopamine agonists, as compared to levodopa, odds ratios and 95% confidence intervals were calculated through conditional logistic regression. In the cohort of 25,459 Parkinson's disease patients (11,151 new users of dopamine agonists, 14,308 new users of levodopa), 518 incident heart failure cases were identified during follow-up. Compared to levodopa, no increased risk of heart failure was found for ergot dopamine agonists (odds ratio: 1.03; 95% confidence interval: 0.69-1.55). Among non-ergot dopamine agonists, only pramipexole was associated with an increased risk of heart failure (odds ratio: 1.61; 95%confidence interval: 1.09-2.38), especially in the first three months of therapy (odds ratio: 3.06; 95% confidence interval: 1.74-5.39) and in patients aged 80 years and older (odds ratio: 3.30; 95% confidence interval: 1.62-7.13). The results of this study indicate that ergot dopamine agonist use in Parkinson's disease patients was not associated with an increased risk of newly diagnosed heart failure. Among non-ergot dopamine agonists, we observed a statistically significant association between pramipexole use and heart failure, especially during the first months of therapy and in very old patients.

Automating classification of free-text electronic health records for epidemiological studies.

PURPOSE: Increasingly, patient information is stored in electronic medical records, which could be reused for research. Often these records comprise unstructured narrative data, which are cumbersome to analyze. The authors investigated whether text mining can make these data suitable for epidemiological studies and compared a concept recognition approach and a range of machine learning techniques that require a manually annotated training set. The authors show how this training set can be created with minimal effort by using a broad database query. METHODS: The approaches were tested on two data sets: a publicly available set of English radiology reports for which International Classification of Diseases, Ninth Revision, Clinical Modification code needed to be assigned and a set of Dutch GP records that needed to be classified as either liver disorder cases or noncases. Performance was tested against a manually created gold standard. RESULTS: The best overall performance was achieved by a combination of a manually created filter for removing negations and speculations and rule learning algorithms such as RIPPER, with high scores on both the radiology reports (positive predictive value = 0.88, sensitivity = 0.85, specificity = 1.00) and the GP records (positive predictive value = 0.89, sensitivity =0.91, specificity =0.76). CONCLUSIONS: Although a training set still needs to be created manually, text mining can help reduce the amount of manual work needed to incorporate narrative data in an epidemiological study and will make the data extraction more reproducible. An advantage of machine learning is that it is able to pick up specific language use, such as abbreviations and synonyms used by physicians.

Suboptimal gastroprotective coverage of NSAID use and the risk of upper gastrointestinal bleeding and ulcers: an observational study using three European databases.

BACKGROUND: Gastro-protective agents (GPA) are co-prescribed with non-steroidal anti-inflammatory drugs (NSAID) to lower the risk of upper gastrointestinal (UGI) events. It is unknown to what extent the protective effect is influenced by therapy adherence. AIM: To study the association between GPA adherence and UGI events among non-selective (ns) NSAID users. METHODS: The General Practice Research Database (UK 1998-2008), the Integrated Primary Care Information database (the Netherlands 1996-2007) and the Health Search/CSD Longitudinal Patient Database (Italy 2000-2007) were used. A nested case-control design was employed within a cohort of nsNSAID users aged ≥50 years, who also used a GPA. UGI event cases (UGI bleeding and/or symptomatic ulcer with/without obstruction/perforation) were matched to event-free members of the cohort for age, sex, database and calendar time. Adherence to GPA was calculated as the proportion of nsNSAID treatment days covered by a GPA prescription. Adjusted OR with 95% CI were calculated. RESULTS: The cohort consisted of 618 684 NSAID users, generating 1 107 266 nsNSAID episodes. Of these, 117 307 (10.6%) were (partly) covered by GPA, 4.9% of which with a GPA coverage <20% (non-adherence), and 68.1% with a GPA coverage >80% (full adherence). 339 patients experienced an event. Among non-adherers, the OR was 2.39 (95% CI 1.66 to 3.44) for all UGI events and 1.89 (95% CI 1.09 to 3.28) for UGI bleeding alone, compared to full adherers. CONCLUSIONS: The risk of UGI events was significantly higher in nsNSAID users with GPA non-adherence. This underlines the importance of strategies to improve GPA adherence.

Guillain-Barré syndrome: background incidence rates in The Netherlands.

Guillain-Barré syndrome (GBS) is a (sub)acute polyradiculoneuropathy, which may occur following immunization. To interpret the occurrence of GBS after introduction of large-scale immunization programmes, it is important to define recent background incidence rates (IRs) of GBS. We used a general practitioner electronic medical record database to assess age-specific GBS IRs between 1996 and 2008 in The Netherlands. All possible GBS cases were manually reviewed. Validated incident cases were reviewed by a neurologist (B. J.) for diagnostic certainty using the GBS case definition of the Brighton Collaboration (BC). In a population of 638,891 persons, we identified 23 validated incident GBS cases (mean age 46 years). IR was 1.14 per 100,000 person years (95% confidence interval [CI] 0.67-1.61) and was lower for people under 50 years (0.76; 95%CI 0.41-1.32) compared with elderly of 50 years or older (1.80; 95%CI 0.98-3.05). Only six cases fulfilled level 1 or 2 of diagnostic certainty of the BC case definition. IR of GBS increases with age. As vaccinations are often targeted at specific age groups, age-specific rates should be used to monitor GBS observed versus expected rates after introduction of large-scale vaccination programmes.

Association between calcium channel blockers and gingival hyperplasia.

AIM: To study the effect of the dose and type of calcium channel blockers (CCBs) on the risk of gingival hyperplasia and to quantify this association. METHODS: The study was conducted within the Integrated Primary Care Information Project in The Netherlands. A nested case-control study was designed within a cohort of all patients who were new users of either CCBs or drugs interacting with the renin-angiotensin system (RAS). Cases were all individuals with a validated diagnosis of gingival hyperplasia. Controls were matched on age, gender and index date. RESULTS: Within the study population, 103 cases of gingival hyperplasia were identified and matched to 7677 controls. The risk of gingival hyperplasia was higher in current users of CCBs [adjusted odds ratio (OR(adj)) 2.2, 95% confidence intervals (95% CI): 1.4-3.4], especially in dihydropyridines (OR(adj) 2.1, 95% CI: 1.3-3.5) and benzothiazepine derivatives (OR(adj) 2.9, 95% CI: 1.3-6.5) than in RAS drug users. The risk increased in patients using more than the recommended daily dose (OR(adj) 3.0, 95% CI: 1.6-5.5) and when the duration of current use was <1 month (OR(adj) 5.2, 95% CI: 2.1-12.6). CONCLUSION: This study shows that the risk of gingival hyperplasia is twofold higher in current users of CCBs than in users of RAS drugs. The association was dose dependent and the highest for dihydropyridines or benzothiazepine derivates.

Proton pump inhibitors and the risk of colorectal cancer.

INTRODUCTION: Proton pump inhibitor (PPI) use is associated with increased serum gastrin levels and bacterial overgrowth, resulting in more toxic bile salt formation. Concern has risen that these factors may increase the risk of developing colorectal neoplasia. AIM: To investigate the association between the use of PPIs and the risk of colorectal cancer. METHODS: A population-based case-control study was conducted within the Dutch Primary Care Information (IPCI) database over the period 1996-2005. Cases with colorectal cancer were matched with up to 20 controls on age, gender, calendar time, and duration of follow-up prior to diagnosis. Cumulative exposure to PPIs was assessed in the 5 yr prior to diagnosis with a 1-yr lag time analysis. We calculated adjusted odds ratios (OR) with 95% confidence intervals (95% CI) using multivariate, conditional logistic regression analysis. RESULTS: Within the source population of 457,024 persons, we identified 595 colorectal cancer cases. The odds of colorectal cancer were not increased among patients ever using PPIs compared with patients who never used PPIs (OR 0.85, 95% CI 0.63-1.16). Also, the use of PPIs for >365 days was not associated with a greater risk of colorectal cancer (OR 0.79, 95% CI 0.44-1.41) compared with nonusers. The odds of colorectal cancer in neither the right nor the left hemicolon were significantly increased in patients using PPIs. CONCLUSION: The present study indicates no association between PPI use and the risk of colorectal cancer. Larger numbers of long-term PPI users are needed to confirm the absence of a risk-increasing effect of long-term PPI exposure.

Risk of cardiac valve regurgitation with dopamine agonist use in Parkinson's disease and hyperprolactinaemia: a multi-country, nested case-control study.

BACKGROUND: There is growing evidence that ergot dopamine agonists may induce cardiac valve regurgitation (CVR) in persons with Parkinson's disease. It is unclear whether the CVR risk is increased with ergot-dopamine agonist use in persons with hyperprolactinaemia, in whom the dose is much lower. OBJECTIVE: The aim of the study was to explore the association between different dopamine agonists and CVR in patients with Parkinson's disease or hyperprolactinaemia. DESIGN: Nested case-control studies conducted separately in cohorts of Parkinson's disease and hyperprolactinaemia patients. Cases were patients who developed newly diagnosed CVR. Controls were CVR-free patients from the same cohorts and were matched to cases by age, sex, database and calendar year. SETTING AND PATIENTS: Study patients were identified from over 4.5 million persons in The Health Improvement Network (THIN; UK), Health Search (Italy), and Integrated Primary Care Information (IPCI; the Netherlands) general practice databases in the years 1996-2007. The Parkinson's disease cohort included new users of dopamine agonists or levodopa, while the hyperprolactinaemia cohort included new users or non-users of dopamine agonists. MAIN OUTCOME MEASURE: Risk of newly diagnosed CVR with dopamine agonist use compared with levodopa use in the Parkinson's disease cohort, and dopamine agonist-naïve patients in the hyperprolactinaemia cohort. RESULTS: In the Parkinson's disease cohort (7893 dopamine agonist users, 11 766 levodopa users), 85 incident CVR cases were identified. Increased CVR risk was observed for ergot dopamine agonists (adjusted OR [OR(adj)] 3.82; 95% CI 2.14, 6.81), but not for non-ergot dopamine agonists (OR(adj) 1.20; 95% CI 0.63, 2.29). In the hyperprolactinaemia cohort (6740 dopamine agonist users and 14 299 dopamine agonist-naïve patients), 37 CVR cases were identified during a mean follow-up of 4.5 years and 3.5 years for new users and non-users of dopamine agonists, respectively. However, no association with ever use of ergot dopamine agonists was observed (OR(adj) 0.47; 95% CI 0.20, 1.19). CONCLUSION: Ergot-derived dopamine agonists are associated with an increased risk of CVR in Parkinson's disease but not in hyperprolactinaemia patients.

Exposure to colorectal examinations before a colorectal cancer diagnosis: a case-control study.

OBJECTIVES: To assess the prior exposure to colorectal examinations between colorectal cancer (CRC) patients and matched control participants to estimate the effect of these examinations on the development of CRC and to obtain insight into the background incidence of colorectal examinations. METHODS: A population-based case-control study was conducted within the Dutch Integrated Primary Care Information database over the period 1996-2005. All incident CRC cases were matched with up to 18 controls (n=7,790) for age, sex, index date (date of CRC diagnosis) and follow-up before diagnosis. All colorectal examinations performed in symptomatic participants in the period 0.5-5 years before index date were considered in the analyses. RESULTS: Within the source population of 457 024 persons, we identified 594 incident cases of CRC. In the period 0.5-5 years before index date 2.9% (17 of 594) of the CRC cases had undergone colorectal examinations, compared with 4.4% (346 of 7790) in the control population [odds ratio (ORadj): 0.56, 95% confidence interval (CI): 0.33-0.94]. For left-sided CRC, significantly more controls than cases had undergone a colorectal examination (4.7 vs. 2.0%, respectively, ORadj: 0.36, 95% CI: 0.17-0.76), which was not seen for right-sided CRCs (3.3 vs. 3.9%, respectively, ORadj: 0.98, 95% CI: 0.42-2.25). CONCLUSION: Patients diagnosed with CRC were less likely than controls to have had a colorectal examination in previous years, being more pronounced in patients diagnosed with left-sided CRCs. If diagnostic examinations have a similar protective effect as screening examinations, this finding supports the concept that colorectal examination can have a major impact on the reduction of CRC risk.

The effect of anticholinergic agents on gastro-oesophageal reflux and related disorders.

The most important risk factor of oesophageal adenocarcinoma is gastro-oesophageal reflux disease. Gastro-oesophageal reflux disease is in itself a common disorder, giving bothersome symptoms. In daily clinical practice, anticholinergic drugs are believed to increase the risk of gastro-oesophageal reflux through effects on the lower oesophageal sphincter. In this review we discuss the available literature on the potential association between the use of drugs with anticholinergic properties and the risk of gastro-oesophageal reflux-related disorders.

Tricyclic antidepressants and the risk of reflux esophagitis.

OBJECTIVE: Incompetence of the lower esophageal sphincter (LES) is a key factor in the pathogenesis of gastroesophageal reflux disease (GERD). Drugs with anticholinergic properties, such as tricyclic antidepressants (TCAs), may facilitate GERD by a relaxing effect on the LES. AIM: To investigate whether the use of TCAs is associated with an increased risk of reflux esophagitis (RE). METHOD: A population-based case-control study was conducted within a large Dutch primary care database over the period 1996-2005. Cases with endoscopy-confirmed RE were identified and matched with up to 10 controls on gender, age, GP practice, and calendar time. Exposure to TCAs was assessed in the year prior to diagnosis and categorized as current (last prescription covered or ended within one month prior to the index date), past, and no use. The relative risk of RE was estimated by odds ratios (OR) with 95% confidence intervals (95% CI) using multivariate conditional logistic regression analysis. RESULTS: During the study period, 1,462 cases with endoscopy-confirmed RE were identified. The risk of RE was increased in current TCA users (OR(adj) 1.61, 95% CI 1.04-2.50). Drug-specific analyses revealed that only clomipramine was associated with an increased risk of RE (OR(adj) 4.6, 95% CI 2.0-10.6) in a duration- and dose-dependent manner (OR(adj) 7.1, 95% CI 2.7-19.2 for use >180 days and OR(adj) 9.2, 95% CI 1.6-51.5 for >1 DDD equivalent/day). CONCLUSION: No association was observed between the risk of RE and the use of TCAs other than clomipramine. The association between RE and clomipramine might be drug-related or a result of the underlying indication.

Protective effect of nonsteroidal anti-inflammatory drugs on colorectal adenomas is modified by a polymorphism in peroxisome proliferator-activated receptor delta.

OBJECTIVE: Nonsteroidal anti-inflammatory drugs (NSAIDs) are associated with a decreased risk of colorectal tumors. Single nucleotide polymorphisms (SNPs) in target genes of NSAID action, and their haplotypes, might modulate this protective effect. METHODS: A case-control study including 724 cases and 682 controls was used to evaluate the effect of NSAIDs on colorectal adenoma risk in The Netherlands, a country in which NSAID use is relatively low. Cases and controls were classified according to presence or absence of endoscopy-proven, pathology-confirmed colorectal adenomas, ever in their lives. Thirteen SNPs in four genes (PPARdelta, PPARgamma, PTGS1 and PTGS2) were genotyped in 787 subjects (384 cases and 403 controls). RESULTS: Compared to non-regular users (< 12 times/year), regular users of NSAIDs (> or = 12 times/year) had a lower risk of colorectal adenomas (odds ratio (OR): 0.75, 95% confidence interval (CI): 0.56-0.99). The results were similar for aspirin only. We found an interaction between SNP c.-789C>T in PPARdelta and NSAID use (P=0.03). The protective effect of NSAIDs was strengthened for regular users with the PPARdelta CT or TT genotypes (OR: 0.35, 95%CI: 0.11-1.13), whereas a positive association was observed for non-regular users with these genotypes (OR: 2.24, 95%CI: 1.06-4.73) as compared to non-regular users with the CC genotype. Also, a statistically significant interaction between a major haplotype containing the minor allele of this SNP and NSAID use was observed. CONCLUSIONS: This study confirms the protective effect of NSAIDs and suggests a modulating effect of a SNP in the promoter of PPARdelta.