Pharmacoepidemiology and Drug Safety最新文献

Purpose: This study extends a version of the high-dimensional propensity score (HDPS) recently modified for the UK electronic health record setting, by enriching primary care data with hospital data. The performance of this modified approach is assessed via the estimation of a well-established association, the reduced risk of upper gastrointestinal bleeding (UGIB) in cyclo-oxygenase-2 inhibitor (COX-2i) users versus non-steroidal anti-inflammatory drug (NSAID) users.

Methods: We conducted an active-comparator, new-user cohort study using UK primary care data from the Clinical Practice Research Datalink GOLD database, with linkages to hospitalisation and mortality records. We included individuals with osteoarthritis initiating NSAIDs or COX-2is between 2000-2004. We used Cox proportional hazards models to estimate the hazard ratio (HR) for UGIB, adjusting for confounders using investigator-specified and HDPS-derived propensity scores. Sensitivity analyses were conducted varying the number of HDPS covariates included and the covariate assessment period.

Results: We identified 74 443 and 25 742 new users of NSAID and COX-2i users, respectively. The unadjusted HR for UGIB comparing COX-2i and NSAID users was 1.28 (95% CI: 0.95-1.72). Of the included HDPS covariates, 26% originated from the hospitalisation dimension, a source not considered in previous applications indicating the considerable information contained in these data on proxies of potential confounders. The modified-HDPS obtained similar results to the other studies, shifting the HR closer to the expected association (HR 0.86; 95% CI: 0.58-1.26).

Conclusion: We demonstrate the ability of the modified-HDPS to obtain similar results to comparable pharmacoepidemiological studies and randomised trials, highlighting the potential benefit of these approaches in UK EHRs more widely and the value of adding hospital data to enrich the pool of covariates available for the HDPS algorithm.

Purpose: To (a) examine longitudinal trends in prescribing of first- and second-line non-insulin pharmacotherapies (NIPs) among reproductive-aged women in Australia between 2013 and 2021 and (b) explore concurrent use of highly effective long-acting reversible contraceptives (LARCs), as well as other hormonal contraceptives, at the time of first dispensing of NIP.

Methods: Using a 10% random sample of Australian women aged 18-44 years from dispensing claims from the Pharmaceutical Benefits Scheme (PBS), the annual prescription of at least one NIP was reported as a rate per 1000 women. Concurrent LARC use was identified where the date of contraceptive supply plus the likely duration of efficacy overlapped with the first dispensing date of NIP.

Results: The overall rate of NIP use has increased from 14.40 to 23.15/1000 women between 2013 and 2021, with increases observed in the rate of women prescribed the first-line agent metformin alone (11.94-18.41/1000), metformin and a second-line NIP (2.17-3.88/1000), and second-line NIP alone (0.29-0.85/1000). When compared with initiating treatment with metformin, the proportion of women considered concurrent LARC users or any contraceptive method was modestly higher for those commencing treatment with a second-line NIP (17.0% vs. 12.7% [aOR, 1.09, 95% CI: 1.02, 1.17] and 26.7% vs. 20.5% [aOR: 1.12, 95% CI: 1.05, 1.19], respectively).

Conclusion: There is increasing use of NIP amongst reproductive-aged women in Australia, with rates of use of second-line NIPs almost doubling between 2013 and 2021. While concurrent use of LARC appears higher among those prescribed second-line NIP, compared with metformin, rates of LARC use still appear low.

Purpose: Our objective was to estimate the effect of initiating an inhaled corticosteroids-containing single-inhaler triple agent (ICS-LABA-LAMA) compared with a single-inhaler LABA-LAMA dual bronchodilator in patients with COPD on the risk of severe COVID-19 prior to the roll-out of vaccines.

Methods: We conducted a cohort study emulating a randomized trial, among patients with COPD aged 40 years or more in the UK, comparing those who initiated a triple inhaler with those who initiated a dual bronchodilator inhaler between March 1 and December 31, 2020. Weighting by fine stratification of the propensity score was used to account for confounders. The risk of hospitalized COVID-19 was compared with a Cox proportional hazards model in an as-treated analysis with a 30-day grace period.

Results: The study cohort included 876 patients initiating a triple inhaler and 5010 initiating a dual LABA-LAMA inhaler. The adjusted incidence rate of hospitalized COVID-19 was 5.6 per 100 person-years in the triple inhaler group and 2.9 per 100 person-years in the dual inhaler group, with a corresponding hazard ratio (HR) of 1.96 (95% confidence interval 1.01-3.77). Sensitivity analyses on the duration of the grace period, using an intent-to-treat exposure classification, or starting follow-up 14 days after treatment initiation (accounting for treatment initiation for an undocumented SARS-CoV-2 infection) were generally consistent with the main analysis.

Conclusions: Patients with COPD prescribed an ICS-containing triple inhaler were potentially exposed to an increased risk of severe COVID-19 prior to the vaccine era. As SARS-CoV-2 continues to cause significant burden, these findings should be considered when determining initiation of inhaled treatment in COPD.

Aims: Randomized controlled trials (RCTs) have high internal validity but often have limited generalizability. To our knowledge, there have been no previous studies emulating RCTs using real-world data to evaluate the risk of major adverse cardiovascular events (MACE) among patients with type 2 diabetes mellitus (T2DM) treated with long-acting insulin analogues.

Methods: We emulated the DEVOTE trial of insulin degludec vs. glargine among patients with T2DM using data from the United Kingdom's Clinical Practice Research Datalink. DEVOTE eligible and ineligible subpopulations were created. Cox proportional hazards models with inverse probability of treatment weighting were used to estimate hazard ratios (HRs) and corresponding confidence intervals (CIs) for MACE comparing new users of insulin degludec to new users of insulin glargine overall and in the eligible/ineligible subpopulations.

Results: There were 10 430 patients in the overall population, 5280 in the DEVOTE eligible population, and 5150 in the DEVOTE ineligible population. The overall (HR: 1.36, 95% CI: 0.83, 1.86) and DEVOTE eligible populations (HR: 1.07, 95% CI: 0.63, 1.58) were compatible with findings from the DEVOTE trial (HR: 0.91, 95% CI: 0.78, 1.06) for the risk of MACE. Due to a low number of events, the DEVOTE ineligible population had deviations in point estimates and wider CIs (HR: 2.19, 95% CI: 0.30, 3.83).

Conclusion: The risk of MACE among patients with T2DM newly prescribed insulin degludec compared to insulin glargine was consistent between the overall population and the DEVOTE eligible subpopulation, while the DEVOTE ineligible population had discrepant point estimates.

Purpose: Real-world evidence (RWE) addresses clinical trial limitations by capturing more representative patient populations and improves evaluation of anticancer treatments, although it becomes available only years after market authorization. As many RWE sources capture only parts of the healthcare continuum, dataset linkage is necessary to enhance data richness. Linkage quality must be assessed to prevent information bias due to incomplete data linkage.

Methods: We evaluated diagnosis concordance for lung cancer (LC), melanoma, and renal cell cancer (RCC) in England. Patients were matched based on national health service (NHS) number, sex and date of birth. Eligible patients were drawn from the National Cancer Registry and Analysis Service (NCRAS), and matched with three other datasets: Clinical Research Practice Database Aurum (CPRD Aurum), Hospital Episode Statistics Admitted Patient Care (HES-APC), and systemic anticancer treatment (SACT). Concordance was evaluated for cancer diagnosis and date of diagnosis. Determinants of non-concordance were investigated to assess representativeness.

Results: In total, 89 797 patients with LC, melanoma or RCC were identified, and concordance of cancer diagnosis records between NCRAS, CPRD Aurum and HES-APC exceeded 70%. Because patients are only registered in SACT upon receiving systemic anticancer treatment, matched numbers in SACT were significantly lower (3.0%-21.1%), as anticipated, particularly among patients over 80 years of age. However, differences in patient characteristics across datasets were limited. Concordance analyses showed that the majority of cases with LC diagnoses were registered within 3 months of the initial diagnosis within all data sources, whereas melanoma and RCC showed longer delays.

Conclusions: Given the high concordance, NCRAS data can be enriched with HES-APC and CPRD Aurum, and further complemented by SACT for systemic therapy. Provided that SACT undergoes further validation, linkage between NCRAS, CPRD Aurum, HES-APC, and SACT may be a promising resource for RWE generation in oncology research.

Purpose: To determine the positive predictive values (PPVs) of ICD-9- and ICD-10-based diagnostic coding algorithms to identify periprosthetic joint infection (PJI) following total hip arthroplasty (THA) within the United States (US) Veterans Health Administration (VHA).

Methods: We selected patients with: (1) any position hospital discharge ICD-9 or ICD-10 diagnosis of PJI, (2) ICD-9, ICD-10, or current procedural terminology (CPT) procedure codes for THA any time prior to PJI diagnosis, (3) CPT code for hip X-ray within ±90 days of the PJI diagnosis, and (4) 1 or more CPT codes for arthrocentesis, arthrotomy, or revision arthroplasty all occurring within ±90 days of the PJI diagnosis date. We obtained separate samples of patients for ICD-9 and ICD-10-based PJI diagnoses. These samples were stratified by THA medical center volume. Infectious disease physicians adjudicated each identified PJI event. The PPV (95% confidence interval [CI]) for the ICD-9 and ICD-10 PJI algorithms were calculated.

Results: Among the 90 sampled hip PJI events for the ICD-9 era, 79 were confirmed PJIs (PPV 87.8%, 95% CI 79.2%-93.7%). For the 90 sampled hip PJI events for the ICD-10 era, 72 were confirmed PJIs (PPV 80.0%, 95% CI 70.3%-87.7%).

Conclusion: These algorithms yielded a PPV of 87.8% (ICD-9) and 80.0% (ICD-10), for confirmed PJI events and could be considered for use in future pharmacoepidemiologic studies.

Aim: Preclinical studies have suggested that glucagon-like peptide-1 receptor agonists (GLP-1RAs) may induce thyroid gland hyperplasia, raising concerns about potential thyroid-related risks. Given their increasing use and the limited evidence on thyroid safety, this study assessed the association between GLP-1RA use and the incidence of hypothyroidism using real-world data.

Methods: We conducted an active-comparator, new-user cohort study using the Real-World Evidence Research Network (SRWEN) (2016-2023). Adults (≥ 18 years) initiating GLP-1RAs or dipeptidyl peptidase-4 inhibitors (DPP-4is) were followed from first prescription until hypothyroidism, treatment discontinuation, switching, death, or study end. Hypothyroidism was identified through ICD-10 codes or levothyroxine prescriptions. Inverse probability of treatment weighting was applied to adjust for confounding, and weighted Cox proportional hazards models were used to estimate hazard ratios (HRs). RStudio 4.4.0 was used for analyses.

Results: A total of 47 017 patients were included (6800 GLP-1RA users; 40 217 DPP-4i users). GLP-1RA users were younger (mean age 50 vs. 58 years) and more often female. The incidence rate of hypothyroidism was 128 per 10 000 person-years in GLP-1RA users compared to 150 per 10 000 person-years in DPP-4i users. GLP-1RA use was not associated with a statistically significant risk of hypothyroidism (adjusted HR 1.04, 95% CI 0.69-1.57). Sensitivity analyses extending follow-up by 30 and 60 days yielded consistent findings.

Conclusion: In this real-world analysis, GLP-1RA use was not associated with an increased incidence of hypothyroidism compared to DPP-4is. Findings were consistent across sensitivity and subgroup analyses. Although findings do not suggest a short-term risk, longer-term studies are warranted to further evaluate thyroid safety.

Purpose: Ascertaining migraine in electronic healthcare data is challenging because of likely diagnosis underrecording and treatment with over-the-counter analgesics, which cannot be used as disease proxies. Algorithm-identified migraine prevalence may depend on algorithm characteristics and target population.

Methods: To describe migraine-identifying algorithms implemented in electronic healthcare data sources and summarize validation results and observed migraine prevalence, we searched PubMed for peer-reviewed, English-language, original research articles that identified migraine in adults using electronic algorithms in electronic healthcare data. We summarized algorithms, validation results, and migraine prevalence (PROSPERO: CRD42023491279).

Results: Of 360 unique titles and abstracts, 50 articles (14%) were selected for full-text review; of them, 41 articles (82%) were finally included: 16 were studies conducted in Europe, 13 in North America, and 12 in Asia. Sixteen studies (39%) identified migraine only using diagnosis codes, 5 (12%) only treatments, 9 (22%) diagnosis and/or treatment codes, and 11 (27%) diagnosis codes, treatments, and setting (e.g., primary care, specialist consultation). Reported migraine prevalence in the general population ranged between 4% and 17%. Only two studies reported validation results: one identified prevention-eligible patients with migraine (positive predictive value [PPV] = 97%), and one identified migraine on the basis of calculated probabilities with PPVs between 74% and 92%.

Conclusion: Finding patients with migraine is feasible in various types of data sources; preferred algorithms vary; algorithm performance is mostly unknown. Identifying chronic migraine or other complex types of migraine requires combining diagnosis codes, treatments, and care settings, which is possible in only some data sources.

Purpose: To evaluate trends and disparities in herpes zoster vaccination among US older adults with diabetes.

Methods: Data from the 2008 to 2023 National Health Interview Survey were used. Joinpoint regression analysis was performed to analyze trends in herpes zoster vaccination. A multivariable logistic regression model was used to identify factors associated with herpes zoster vaccination.

Results: A total of 42 377 participants with diabetes were included, representing approximately 18 million US older adults with diabetes. From 2008 to 2023, the prevalence of herpes zoster vaccination increased nearly tenfold, from 4.2% in 2008 to 42.2% in 2023 (average annual percent change = 14.09, p < 0.01), with similar overall trends observed in adults without diabetes (p = 0.08). Upward trends were also observed across age groups and diabetes types. Several factors, including age, race/ethnicity, region, educational level, health insurance, income, perceived health status, flu and pneumococcal vaccination, comorbid atherosclerotic cardiovascular disease and cancer, were associated with herpes zoster vaccination.

Conclusion: Herpes zoster vaccine coverage has surged among US older adults with diabetes over the past 16 years. However, disparities in vaccination remain, underscoring the need for targeted policies and interventions to improve coverage.

Purpose: To identify the best method for selecting index dates when constructing external comparator arms (ECAs) from real-world data for comparison with single-arm trials (SATs).

Methods: We evaluated four approaches for index date selection-first eligible line, last eligible line, random eligible line, and all eligible lines-using causal inference reasoning, numerical examples and a simulation study. Simulations modeled survival across multiple lines of therapy under scenarios with varying eligibility patterns and treatment effects. Overall survival (OS) estimates comparing SAT and ECA populations were obtained using stratified Cox models and propensity-weighted Cox models, adjusted for line of therapy and patient state.

Results: Including all eligible lines produced unbiased OS estimates across scenarios. Selecting the last eligible line introduced substantial bias, while random selection led to moderate bias.

Conclusions: Using all eligible lines of therapy for each patient when constructing ECAs minimizes bias and preserves the SAT target population. Alternative methods can lead to biased estimates or, in the case of the first eligible line method, require changes to the clinical question that may shrink the SAT population. We recommend adopting the all eligible lines method with variance correction and adjustment for line of therapy to ensure valid comparative effectiveness analyses.