Pharmacoepidemiology and Drug Safety最新文献

Background: Despite testing of epidemiological methods in US Claims databases for signal detection, such data sources have not become a routine capability. The Self Controlled Case Series (SCCS) is one of the most promising methods for drug safety signal detection using Real World Data, and incorporating active comparators could potentially improve its performance by addressing confounding by indication.

Objectives: This study aims to evaluate the performance of the SCCS with and without active comparators for signal detection using US Merative MarketScan Commercial Claims and Medicare databases.

Methods: We applied the SCCS to macrolide and fluoroquinolone antibiotics, using amoxicillin and cefalexin as active comparators. In total, 7 drugs and 30 outcomes from all organ classes were selected. We developed a reference set of 104 positive controls and 58 negative controls, using a taxonomy framework to ensure the selected drug outcome pairs are theoretically well suited to the SCCS design. A two-year observation period with a 30-day risk window after each dispensing was used. Diagnostic performance was measured using sensitivity and specificity with respect to the product labels.

Results: The SCCS without active comparators achieved sensitivities of 0.73 and 0.72 and specificities of 0.68 and 0.62 in commercial and Medicare claims, respectively, for pairs with sufficient power. Active comparators increased specificity up to 0.84 and 0.86, respectively, in Commercial Claims and Medicare but decreased sensitivity to 0.45 and 0.36.

Conclusions: MarketScan databases are potentially suitable for drug safety signal detection due to their large size and information contained. Using a carefully designed reference set of drug-outcome pairs well suited to the study design, the SCCS, while imperfect, performed comparably to optimal settings identified in previously published studies. Active comparators did not enhance overall performance but showed improved specificity by better controlling confounding by indication at the cost of reduced sensitivity.

Purpose: Combined oral contraceptives (COCs) are contraindicated in migraine with aura due to stroke risk, and some hormone therapy for menopause guidelines recommend caution in this population. However, this guidance is informed by sparse or older evidence reflective of higher doses than typically prescribed today. This study aimed to describe modern-day utilization of COCs and hormone therapy among female individuals with migraine with aura from 2000 to 2024.

Methods: Using United Kingdom medical record data, this study evaluated the use of COCs and progestogen-only pills (POPs) among reproductive-age individuals and hormone therapy among post-reproductive age individuals, before and after migraine with aura diagnosis. Post-diagnosis medication utilization was described relative to baseline characteristics and pre-diagnosis use of each medication, overall and longitudinally.

Results: Among 142 867 individuals of reproductive age, 84 374 (59%) used oral contraceptives on or after migraine with aura diagnosis, predominantly POPs (n = 75 823, 53% of those with any oral contraceptive) over COCs (n = 21 968, 15%). Most oral contraceptive users in the year pre-diagnosis used COCs (n = 36 909/56760, 66%). Among 46 913 individuals of post-reproductive age, 20 990 (45%) had a prescription for hormone therapy after migraine with aura diagnosis, with transdermal formulations used increasingly over calendar time.

Conclusions: Utilization of COCs declined but did not fully cease after migraine with aura diagnosis. Post-diagnosis utilization of hormone therapy for menopause was common. Given this utilization among individuals with migraine with aura, high-quality evidence quantifying the risk of stroke associated with modern-day use of these medications in this population is needed to inform patient and provider decision making.

Purpose: Pharmacotherapy during pregnancy should be approached with caution due to the potential risk of adverse effects, including birth defects, in the fetus. Appropriate post-marketing surveillance and perinatal pharmacoepidemiology are essential to ensure the safety of pharmacotherapy during pregnancy. However, due to limited research infrastructure in pharmacoepidemiology, collecting reliable data on drug safety in pregnant women and infants remains a challenge in Japan. Thus, we examined the suitability (fitness for purpose) of Japanese medical databases for perinatal studies to establish infrastructure in this field.

Methods: We assessed seven available databases in Japan: the DeSC database, EBM provider, Japanese Adverse Drug Event Report database, JMDC claims database, MDV analyzer, the National Database of Health Insurance Claims, and Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study). To assess the quality of databases for maternal studies, we evaluated the content of each mother-infant linkable database based on the core data elements recommended for perinatal pharmacoepidemiology.

Results: Three of the databases were found to be mother-infant linkable: the DeSC database, JMDC claims database, and the TMM BirThree Cohort Study. The coverage of core data elements in perinatal pharmacoepidemiology was 73.5% in the DeSC and JMDC claims databases and 92.9% in the TMM BirThree Cohort Study.

Conclusion: Some representative medical databases in Japan are well suited for use in perinatal pharmacoepidemiologic research on infant outcomes.

Key changes in the pharmacoepidemiological research environment had a significant influence on the activities of the European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP) over the last decade. These changes included the SARS-CoV-2 pandemic, the increased access to anonymized real-world data (RWD) sources, the integration of real-world evidence (RWE) into regulatory and public health decision-making, and the emergence of new technologies and methods. This paper describes how ENCePP has evolved in this changing environment to strengthen pharmacoepidemiological methods and practice in Europe and globally. It also provides future perspectives for the network. Through a collaborative approach in non-interventional research, ENCePP will collectively continue to promote excellence for RWE generation, supporting the safe and effective use of medicines.

Background: The self controlled case series (SCCS) is one of the most promising methods for drug safety signal detection using real world data (RWD), and incorporating active comparators could potentially improve its performance by addressing time-varying confounding by indication. The 'Système National des Données de Santé' (SNDS) is a large nationwide administrative claims database, which has not been used extensively for drug safety signal detection. While comparable in size to other RWD sources, it is unclear to what extent the performance of SCCS correlates with that in other sources.

Objectives: This study aims to evaluate the performance of the SCCS with and without active comparators for signal detection in the French administrative healthcare database SNDS.

Methods: We applied the SCCS to macrolide and fluoroquinolone antibiotics, using amoxicillin as the active comparator. Amoxicillin was chosen as an active comparator with similar indications. In total, 7 drugs and 30 outcomes from all organ classes were selected. We developed a reference set of 104 positive controls and 58 negative controls, using a taxonomy framework to ensure the selected drug outcome pairs are theoretically well suited to the SCCS design. The observation period lasted 2 years, with a 30-day risk window after each dispensing. Diagnostic performance was measured using sensitivity and specificity with respect to the product labels.

Results: The sensitivity and specificity of the SCCS without active comparator were 0.89 and 0.43, respectively, when limited to pairs with satisfactory power. Specificity increased up to 0.91 with active comparators; however, sensitivity decreased to 0.52.

Conclusions: The SNDS is a useful data source for signal detection, particularly for outcomes captured in hospitals. Using a carefully designed reference set of drug-outcome pairs well suited to the study design, the SCCS achieved satisfactory performance for signal detection in this database. In this study, the use of active comparators improved overall performance at the expense of greatly reduced sensitivity.

Purpose: This study aims to compare the effectiveness and safety of standard and reduced initial doses of regorafenib in patients with metastatic colorectal cancer (mCRC).

Materials and methods: This retrospective observational study was conducted using the Taipei Medical University Clinical Research Database. Patients aged 20 years or older who received regorafenib for mCRC between January 2014 and December 2021 were included. Patients were divided into standard initial dose (160 mg) and reduced initial dose (< 160 mg) groups. Time-to-treatment discontinuation (TTD), overall survival (OS), and the incidence of five common adverse events were compared between groups.

Results: Among 266 patients, 58 received the standard initial dose and 208 received the reduced initial dose; the median TTD was 68.0 days and 64.5 days, respectively (p = 0.25). The median OS was 9.7 months for the standard-dose group and 6.7 months for the reduced-dose group (p = 0.01). In the multivariate Cox analysis, the reduced initial dose was associated with shorter survival (hazard ratio 1.66 [95% confidence interval 1.22-2.30]). Hand-foot skin reaction and total bilirubin elevation were less common in the reduced-dose group (p = 0.01 and 0.03, respectively). Excluding concurrent anti-cancer drug users led to a similar median OS between the two dosing groups (p = 0.12).

Conclusion: No difference in TTD was observed between the dosing groups. The reduced-dose group had a shorter OS but fewer adverse events. For patients who can tolerate standard doses of regorafenib, a combination of regorafenib and other anti-cancer drugs may be beneficial but would require further investigation.

Purpose: Leukotriene receptor antagonists (LTRAs) are widely prescribed as controller medications for pediatric asthma. However, there have been increasing concerns about potential neuropsychiatric adverse reactions associated with LTRAs. Findings from observational studies have been inconsistent, and direct comparisons of the risk of neuropsychiatric events (NPEs) between LTRAs and inhaled corticosteroids (ICS) remain limited in the pediatric population.

Methods: A retrospective cohort study was conducted utilizing a nationwide claims database (January 2018-April 2022) and a multicenter electronic health record (EHR) database (January 2006-March 2022) from South Korea. Patients aged 5-18 years diagnosed with asthma before initiating LTRA or ICS were included. The primary outcome was NPEs within 90 days of exposure, defined using two methods: diagnostic code-based analysis and natural language processing (NLP)-based analysis using clinical notes. After propensity score stratification, Cox proportional hazards models were used to estimate risks.

Results: The diagnostic code-based analysis on the claims database included 169 636 LTRA users and 28 845 ICS users. There was no statistically significant difference in the risk of NPEs between LTRA and ICS (calibrated hazard ratios [HRs], 1.14 [95% CI, 0.92-1.42]). In the NLP-based analysis using EHR database, 1641 LTRA users and 1607 ICS users were included. The results were consistent with those of the diagnostic code-based analysis (calibrated HR, 1.33 [95% CI, 0.66-2.68]).

Conclusions: LTRA use was not found to be associated with a significantly increased risk of NPEs in children with asthma. These findings offer valuable insights to support clinical decision-making in pediatric asthma treatment.

Purpose: The use of common data models (CDMs) is increasing; however, the complexity of many CDM frameworks constitutes a barrier for their use. For many local and collaborative use cases, simpler CDMs can suffice. Here, we propose the EpiCore CDM, a simple CDM framework for use in Scandinavian pharmacoepidemiological studies.

Methods: The EpiCore CDM was developed based on a set of guiding principles. It should (i) accommodate the most common elements of typical data sources in the field and region, (ii) be accessible to users without needing advanced technical expertise or database infrastructure, (iii) prioritize structural and syntactic harmonization of data and defer clinical concept mapping to the analytical phase, (iv) be usable in both collaborative and single site settings, and (v) include support for quality control procedures.

Results: The EpiCore CDM comprises two mandatory administrative tables (person and observation), six optional event tables (diagnosis, procedure, encounter, drug, primcare, and cancer) and three optional lookup tables (drug_info, organisation_info, and prescriber_info). Each table, along with its columns and constraints is specified according to an EpiCore CDM specification template. This provides easy documentation and integrates with an R-package called EpiCoreAssistant, which provides quality control tools for testing the compliance of a CDM instance with the EpiCore specification. In the event that a project requires customization of the CDM, this is easily implemented in the template and testing. A step-by-step description is presented, demonstrating the steps involved in a typical CDM-based collaborative pharmacoepidemiological study using the EpiCore CDM.

Conclusions: We present the EpiCore CDM, a specification template and an R package that offers a simple framework for improved workflows, standardizations and collaboration, focused on Scandinavian pharmacoepidemiological studies and with relevance for a broad palette of register-based health care researchers.

Purpose: To study the effect of tricyclic antidepressant (TCA)/selective serotonin-norepinephrine reuptake inhibitor (SNRI) versus selective serotonin reuptake inhibitor (SSRI) on the risk of urinary tract infection (UTI) among people with multiple sclerosis (MS).

Methods: A case-control study was conducted using data from the UK Clinical Practice Research Datalink Aurum, nested within a cohort of people with MS. Each person with a UTI was matched to ≤ 4 controls on sex, region, age (±2 years), and time since MS diagnosis (±20%). Conditional logistic regression analyses were conducted to calculate adjusted odds ratios (aORs) with 95% confidence intervals (CIs) to compare TCA/SNRI to SSRI use and each antidepressant group to no antidepressant use. ORs were adjusted for smoking status, BMI, comorbidities, and recent drug prescriptions.

Results: Two thousand six hundred and sixty-four cases were matched to 3722 controls. TCA/SNRI versus SSRI use did not increase the risk of UTI (conditional aOR 1.21 [95% CI: 0.84-1.75]). TCA/SNRI use versus no antidepressant use did show an increased risk (conditional aOR [95% CI: 1.43 [1.21-1.69]), but SSRI versus no use did not (conditional aOR 1.15 [0.96-1.37]).

Conclusions: This study showed no increased risk of UTI for TCA/SNRI versus SSRI use among people with MS. However, effect estimates were imprecise due to small sample sizes.