Using real-world evidence for pharmacovigilance and drug safety-related decision making by a resource-limited health authority: 10 years of experience in Taiwan.

PURPOSE: Real-world evidence has become increasingly relevant in regulatory decision making. Compared to large regulatory bodies, the national pharmacovigilance system in Taiwan is still under development, and the aim of this study is to demonstrate how a resource-limited health authority utilizes real-world evidence in decision making. METHODS: We described different sources of real-world data available in Taiwan and illustrated the structural framework that integrates real-world evidence into Taiwan's national pharmacovigilance system. Additionally, we reviewed real-world studies conducted in the past 10 years and provided examples to show how these studies influenced drug safety-related decision making in Taiwan. RESULTS: During the past 10 years, real-world evidence used when making drug safety-related regulatory decisions in Taiwan was mainly generated from nationwide claims databases, but other sources of real-world data, such as national registries and large electronic hospital databases, also became available recently. Different types of real-world evidence, including drug utilization studies, risk evaluation studies, and risk minimization measure evaluation studies, have been used to support regulatory decisions in Taiwan. CONCLUSIONS: Through collaborations between the government and academics, Taiwan has started to integrate real-world evidence into the national pharmacovigilance system. However, future efforts, including linkages between different sources of real-world data and improvements in procedural and methodological practices, are needed to generate more regulatory-quality real-world evidence.

Risk of cutaneous adverse reactions associated with allopurinol or febuxostat in real-world patients: A nationwide study.

AIMS: Allopurinol carries a well-known risk of cutaneous adverse reactions (CARs). Although febuxostat, a xanthine-oxidase inhibitor with different chemical structure, has been considered an alternative to allopurinol, post-marketing case reports of life-threatening febuxostat-related CARs have been reported. We aimed to compare the risk of CARs between allopurinol and febuxostat in real-world settings and to assess the impact of the market entry of febuxostat on allopurinol use and associated CARs. METHODS: A nationwide study was conducted using Taiwan's National Health Insurance Research Database. In the new-user cohort study, patients who received their first prescriptions of allopurinol or febuxostat were included, and Poisson regression was used to estimate the incidence rate ratios (IRRs) of CARs. In the interrupted time series analysis, time series data on new users and incidence rate of CARs were divided into three periods based on the reimbursement scheme of febuxostat in Taiwan, and segmented regression models were used to estimate changes in both the level and trend in each period. RESULTS: We identified 526 cases of CARs with 487 among new users of allopurinol and 39 among new users of febuxostat (incidence rate: 15.37 vs 3.48 per 1000 person-years). Allopurinol was associated with higher risk of CARs (adjusted IRR 5.55, 95% CI [3.97-7.76]), mild CARs (1.86, [1.24-2.81]), severe CARs (16.75, [8.87-31.62]) and fatal CARs (16.18, [5.05-51.83]) than febuxostat. The overall incidence rates of xanthine-oxidase inhibitor-related CARs decreased from 15.28 to 14.28 per 1000 person-years after the initial reimbursement of febuxostat and further decreased to 9.46 after the reimbursement coverage of febuxostat expanded; however, the changes were not statistically significant. CONCLUSION: Febuxostat can be considered an alternative for patients carrying risk factors for allopurinol-related CARs. However, since there were fatal cases of febuxostat-related CARs, the closely monitoring of symptoms of CARs during the initiation of febuxostat is still warranted.

Temporal trends and patterns in carbamazepine use, related severe cutaneous adverse reactions, and HLA-B*15:02 screening: A nationwide study.

OBJECTIVE: After discovering the association between the HLA-B*15:02 allele and carbamazepine-related severe cutaneous adverse reactions (SCARs), particularly in Southeastern Asian populations, clinical strategies to prevent carbamazepine-related SCARs have changed. We aimed to investigate 10-year trends in carbamazepine use and carbamazepine-related SCARs and to examine the patterns and determinants of HLA-B*15:02 screening in Taiwan. METHODS: A nationwide study was performed using Taiwan's National Health Insurance Research Database. In the first part of the study, new users of carbamazepine were included, and those who experienced SCAR-related admissions were further identified. In the second part of the study, recipients of HLA-B*15:02 screening (reimbursed by Taiwan's National Health Insurance since June 2010) were included and multivariate logistic regression was used to explore factors associated with the use of screening. RESULTS: The numbers of new users of carbamazepine and SCAR cases decreased remarkably during the 10-year period (-82.6% and -87.1%, respectively), and the incidence rates of SCARs showed a downward trend after 2011. The screening rate of the HLA-B*15:02 allele increased to 24.9% in 2014. Neurologists (odds ratio 12.33, 95% confidence interval 9.30-16.35), psychiatrists (9.97, 7.31-13.61), and neurosurgeons (3.23, 2.42-4.32) were more likely to perform screening tests than other specialties were. Physicians practicing in medical centers (6.00, 5.51-6.54) were more likely to perform screening tests than those practicing in other hospitals, whereas the screening rates in clinics remained at 0.0% throughout the study period. SIGNIFICANCE: In recent years, the number of carbamazepine-related SCAR cases has decreased substantially in Taiwan. However, only one-fourth of new users of carbamazepine received HLA-B*15:02 screening, and there were considerable disparities in the screening rates across different physician groups. Policymakers should consider solutions to barriers to implementing screening tests in clinical practice and should not neglect the value of other safety communications and regulations to complement the limitations of pharmacogenomic testing.

Risks of maternal cardiopulmonary events associated with ritodrine for tocolysis: A national database linkage study in 1 831 564 pregnant women.

OBJECTIVE: Real-world data on cardiopulmonary events among pregnant women receiving ß-agonist therapy are scarce. In the present study, we aimed to examine the absolute and relative risks of maternal cardiopulmonary events associated with the use of ß-agonist ritodrine during pregnancy. METHODS: By linking Taiwan's National Birth Certificate Application Database with National Health Insurance data, 1 831 564 pregnancies at ≥20 weeks' gestation were identified. Age-standardized incidence rates of cardiopulmonary events among pregnant women exposed to ritodrine were estimated. Nested case-control analyses were conducted to evaluate the relative risk of pulmonary edema, heart failure, and arrhythmia associated with prior ritodrine use. Cases and controls were matched using risk set sampling, and adjusted odds ratios were estimated using conditional logistic regression models. RESULTS: A total of 189 cases of pulmonary edema, 126 cases of heart failure, and 162 cases of arrhythmia were identified (corresponding age-standardized incidence rates: 20.90, 8.35, and 16.63 per 100 000 among pregnant women only exposed to oral ritodrine; 91.28, 36.01, and 14.61 per 100 000 among those ever exposed to intravenous ritodrine). Exposure to oral ritodrine was associated with a lower increased risk of pulmonary edema (aOR 1.76; 95% CI: 1.12-2.76) and arrhythmia (2.21; 1.47-3.32) whereas exposure to ritodrine injection was associated with a significantly higher risk of pulmonary edema (10.56; 6.39-17.45), arrhythmia (4.15; 1.99-8.64), and heart failure (5.58; 2.27-13.74). CONCLUSIONS: Pregnant women receiving intravenous ritodrine therapy had higher cardiopulmonary risks and should be intensively monitored. While the relative risk associated with oral ritodrine is not pronounced, it should be used judiciously among pregnant women as well.

A New Drug Safety Signal Detection and Triage System Integrating Sequence Symmetry Analysis and Tree-Based Scan Statistics with Longitudinal Data.

Purpose: Development and evaluation of a drug-safety signal detection system integrating data-mining tools in longitudinal data is essential. This study aimed to construct a new triage system using longitudinal data for drug-safety signal detection, integrating data-mining tools, and evaluate adaptability of such system. Patients and Methods: Based on relevant guidelines and structural frameworks in Taiwan's pharmacovigilance system, we constructed a triage system integrating sequence symmetry analysis (SSA) and tree-based scan statistics (TreeScan) as data-mining tools for detecting safety signals. We conducted an exploratory analysis utilizing Taiwan's National Health Insurance Database and selecting two drug classes (sodium-glucose co-transporter-2 inhibitors (SGLT2i) and non-fluorinated quinolones (NFQ)) as chronic and episodic treatment respectively, as examples to test feasibility of the system. Results: Under the proposed system, either cohort-based or self-controlled mining with SSA and TreeScan was selected, based on whether the screened drug had an appropriate comparator. All detected alerts were further classified as known adverse drug reactions (ADRs), events related to other causes or potential signals from the triage algorithm, building on existing drug labels and clinical judgement. Exploratory analysis revealed greater numbers of signals for NFQ with a relatively low proportion of known ADRs; most were related to indication, patient characteristics or bias. No safety signals were found. By contrast, most SGLT2i signals were known ADRs or events related to patient characteristics. Four were potential signals warranting further investigation. Conclusion: The proposed system facilitated active and systematic screening to detect and classify potential safety signals. Countries with real-world longitudinal data could adopt it to streamline drug-safety surveillance.

Risk of myocarditis and pericarditis following coronavirus disease 2019 messenger RNA Vaccination-A nationwide study.

BACKGROUND: An extended interval between the two primary doses may reduce the risk of myocarditis/pericarditis after COVID-19 mRNA vaccination. Taiwan has implemented a two-dose regimen with a 12-week interval for adolescents. Here we present nationwide data of myocarditis/pericarditis following COVID-19 vaccinations. METHODS: Data on adverse events of myocarditis/pericarditis were from the Taiwan Vaccine Adverse Events Reporting System between March 22, 2021, and February 9, 2022. The reporting rates according to sex, age, and vaccine type were calculated. We investigated the rates among young individuals under different two-dose intervals and among those who received two doses of different vaccines. RESULTS: Among 204 cases who met the case definition of myocarditis/pericarditis, 75 cases occurred after the first dose and 129 after the second. The rate of myocarditis/pericarditis after COVID-19 vaccination varied across sex and age groups and was highest after the second dose in males aged 12-17 years (126.79 cases per million vaccinees) for the BNT162b2 vaccine and in males aged 18-24 years (93.84 cases per million vaccinees) for the mRNA-1273 vaccine. The data did not suggest an association between longer between-dose interval and lower rate of myocarditis/pericarditis among males and females aged 18-24 or 25-29 years who received two doses of the BNT162b2 or mRNA-1273 vaccine. Rates of myocarditis/pericarditis in males and females aged 18-49 years after receiving ChAdOx1-S - mRNA-1273 vaccination was significantly higher than after ChAdOx1-S - ChAdOx1-S vaccination. CONCLUSIONS: Myocarditis and pericarditis are rare following mRNA vaccination, with higher risk occurring in young males after the second dose.

Positive Predictive Value of ICD-10 Codes for Cerebral Venous Sinus Thrombosis in Taiwan's National Health Insurance Claims Database.

OBJECTIVE: This study aims to determine the positive predictive value (PPV) of case definitions for cerebral venous sinus thrombosis (CVST) in Taiwan's National Health Insurance claims database based on the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnostic codes. STUDY DESIGN AND SETTING: Inpatient records with ICD-10-CM codes of G08, I629, I636, or I676 were retrieved from the claims data of all hospital branches of Chang Gung Medical Foundation. Manual review of the medical records and images was performed in order to ascertain the diagnosis. The PPV of various case definitions for CVST was estimated. RESULTS: Of the 380 hospitalizations, 166 and 214 were determined to be true-positive and false-positive episodes of acute CVST, respectively. The PPV of the ICD-10-CM codes of G08, I629, I636, and I676 was 88.2%, 2.0%, 100.0%, and 91.3%, respectively. The PPV generally increased when acute CVST was defined as a primary diagnosis or as ICD-10-CM codes plus anticoagulant use. Miscoding in other conditions, tentative diagnosis, and remote episode of CVST were determined as the main reasons for false-positive diagnosis of acute CVST. CONCLUSION: This study determined the PPV of ICD-10-CM codes for identifying CVST, which may offer a reference for future claims-based research.

Association between 9-month isoniazid prophylaxis of latent tuberculosis and severe hepatitis in patients treated with TNF inhibitors.

To investigate associations between isoniazid for latent tuberculosis and risk of severe hepatitis, affecting patients with rheumatoid arthritis or ankylosing spondylitis whose treatment includes tumor necrosis factor inhibitors. Our self-controlled case series study analyzed Taiwan's National Health Insurance Database from 2003 to 2015 to identify RA or AS patients, aged ≥ 20 years, receiving TNF inhibitors and a 9-month single isoniazid treatment. The outcome of interest was hospitalization due to severe hepatitis. We defined risk periods by isoniazid exposure (days): 1-28, 29-56, 57-84, 85-168, 169-252, and 253-280. To compare risk of severe hepatitis in exposed and non-exposed periods, we performed conditional Poisson regressions to generate incidence rate ratios (IRR) and 95% confidence intervals, with adjustment of patients' baseline covariates including age, sex, HBV, HCV and related medication. Of 54,267 RA patients and 137,889 AS patients identified between 2000 and 2015, 11,221 (20.7%) RA and 4,208 (3.1%) AS patients underwent TNFi therapy, with 722 (5%) receiving isoniazid for latent tuberculosis. We identified 31 incident cases (4.3%) of hospitalization due to severe hepatitis. Of these hospitalization events, 5 occurred in the exposed periods, 25 occurred in the INH unexposed periods, and 1 occurred in the pre-exposure period. Compared with non-exposure, the risk of severe hepatitis was higher in exposed periods (incidence rate ratio [IRR]: 5.1, 95% CI: 1.57-16.55), especially 57-84 days (IRR: 17.29, 95% CI: 3.11-96.25) and 85-168 days (IRR:10.55, 95% CI: 1.90-58.51). The INH related fatal hepatotoxicity was not identified in our study. Our findings suggest an association between risk of severe hepatitis and exposure to isoniazid in patients with RA or AS under TNFi therapy, particularly within the exposed period 57-168 days. A close monitoring of liver function is mandatory to minimize the risk, especially within the first 6 months after initiation of 9 months isoniazid.

Trends in reporting drug-associated liver injuries in Taiwan: a focus on amiodarone.

Background A pharmacovigilance database of real-world adverse drug reaction (ADR) reports is helpful for characterising adverse events and identifying new signals after drug approval. Objective This study aimed to analyse trends of ADR reporting in relation to liver injury and to delineate critical factors for suspected drug-related hepatotoxicity with a focus on reports associated with amiodarone. Setting The 2000-2014 Taiwan pharmacovigilance database. Method Relevant Standardized Medical Dictionary for Regulatory Activities queries were used to identify reports associated with liver injury. Information on ADR, patient characteristics, and the verbatim pertaining to amiodarone prescriptions, liver injury, comedications, and comorbidities were extracted and evaluated. Group comparisons between Hy's Law cases and Temple's Corollary cases of suspected amiodarone-related hepatotoxicity were performed. Main outcome measure Number and nature of drug-related liver injuries, particularly those associated with amiodarone. Results Of the 98,777 ADR reports over a 15-year period, 4261 (4.3%) were related to liver injury. Sixty-eight reports contained amiodarone prescriptions, but only 49 (1.1%) were eligible for further analysis. Hepatotoxic cases associated with amiodarone mostly occurred within 1 week, exhibited a hepatocellular pattern, and were more common among elderly individuals. Among 23 discernible cases, four (17.4%) recovered fully from liver injury. The critical Hy's Law cases were associated with shorter height, lower body surface area, and higher average daily doses. Conclusion This study substantiates the importance of ADR reporting. Data pertaining to drug-associated liver injury and factors associated with suspected amiodarone-related hepatotoxicity warrants continual attention in pharmacovigilance for those at risk, especially the elderly.

Impact of Safety-Related Regulations on Codeine Use in Children: A Quasi-Experimental Study Using Taiwan's National Health Insurance Research Database.

INTRODUCTION: Safety concerns regarding potential life-threatening adverse events associated with codeine have resulted in policy decisions to restrict its use in pediatrics. However, whether these drug safety communications have had an immediate and strong impact on codeine use remains in question. OBJECTIVE: We aimed to investigate the impact of the two implemented safety-related regulations (label changes and reimbursement regulations) on the use of codeine for upper respiratory infection (URI) or cough. METHODS: A quasi-experimental study was performed using Taiwan's National Health Insurance Research Database. Quarterly data of codeine prescription rates for URI/cough visits were reported, and an interrupted time series design was used to assess the impact of the safety regulations on the uses of codeine among children with URI/cough visits. Multivariable logistic regression models were used to explore patient and provider characteristics associated with the use of codeine. RESULTS: The safety-related regulations were associated with a significant reduction in codeine prescription rates of -4.24% (95% confidence interval [CI] -4.78 to -3.70), and the relative reduction compared with predicted rates based on preregulation projections was 60.4, 56.6, and 53.2% in the first, second, and third year after the regulations began, respectively. In the postregulation period, physicians specializing in otolaryngology (odds ratio [OR] 1.47, 95% CI 1.45-1.49), practicing in district hospitals (OR 6.84, 95% CI 5.82-8.04) or clinics (OR 6.50, 95% CI 5.54-7.62), and practicing in the least urbanized areas (OR 1.60, 95% CI 1.55-1.64) were more likely to prescribe codeine to children than their counterparts. CONCLUSIONS: Our study provides a successful example of how to effectively reduce the codeine prescriptions in children in the 'real-world' settings, and highlights areas where future effort could be made to improve the safety use of codeine. Future research is warranted to explore whether there was a simultaneous decrease in the incidence rates of codeine-related adverse events following the safety-related regulations.

The reporting completeness of a passive safety surveillance system for pandemic (H1N1) 2009 vaccines: a capture-recapture analysis.

Adverse events following pandemic (H1N1) 2009 vaccines ("2009 H1N1 vaccines") in Taiwan were passively reported to the National Adverse Drug Reaction Reporting System. To evaluate the completeness of spontaneous reporting, cases of death, Guillain-Barré syndrome (GBS), convulsion, Bell's palsy, and idiopathic thrombocytopenic purpura (ITP) after 2009 H1N1 vaccination that occurred between November 1, 2009 and August 31, 2010 were selected from the National Adverse Drug Reaction Reporting System (NADRRS) database and an additionally constructed nationwide large-linked database (LLDB), and matched on a unique personal identifier, date of vaccination (within ±7 days), and date of diagnosis (within ±7 days). Overall, matches occurred between the two data sources included 21 for death, 5 for GBS, 19 for convulsion, 22 for Bell's palsy, and 5 for ITP. The Chapman capture-recapture estimated spontaneous reporting completeness within 0-42 days of vaccination was 4% for death, 71% for GBS, 3% for convulsion, 9% for Bell's palsy, and 15% for ITP. For the interval ≥43 days after vaccination, reporting completeness was 0.1% for death, 14% for GBS, 0.1% for convulsion, <0.1% for Bell's palsy, and 0% for ITP. The estimated-to-expected ratio for Bell's palsy in the interval 0-42 days after vaccination was 1.48 (95% CI 1.11-1.98). Reporting completeness was higher for GBS than other adverse events after 2009 H1N1 vaccination. Linking the NADRRS to existing data sources in a capture-recapture analysis can be considered as an alternative to enhance Taiwan's postlicensure safety assessment of other routine vaccines. Nevertheless, the possibility of an increased risk for Bell's palsy detected by capture-recapture analyses needs further evaluation by controlled studies.

Adverse events following pandemic A (H1N1) 2009 monovalent vaccines in pregnant women--Taiwan, November 2009-August 2010.

BACKGROUND: During the 2009 H1N1 pandemic, pregnant women were prioritized to receive the unadjuvanted or MF59®-adjuvanted pandemic A (H1N1) 2009 monovalent vaccines ("2009 H1N1 vaccines") in Taiwan regardless of stage of pregnancy. Monitoring adverse events following 2009 H1N1 vaccination in pregnant women was a priority for the mass immunization campaign beginning November 2009. METHODS/FINDINGS: We characterized reports to the national passive surveillance from November 2009 through August 2010 involving adverse events following 2009 H1N1 vaccines among pregnant women. Reports from the passive surveillance were matched to a large-linked database on a unique identifier, date of vaccination, and date of diagnosis in a capture-recapture analysis to estimate the true number of spontaneous abortion after 2009 H1N1 vaccination. We verified 16 spontaneous abortions, 11 stillbirths, 4 neonatal deaths, 4 nonpregnancy-specific adverse events, and 2 inadvertent immunizations in recipients who were unaware of pregnancy at time of vaccination. The Chapman capture-recapture estimator of true number of spontaneous abortion after 2009 H1N1 vaccination was 329 (95% confidence interval [CI] 196-553). Of the 14,474 pregnant women who received the 2009 H1N1 vaccines, the estimated risk of spontaneous abortion was 2.3 (95% CI, 1.4-3.8) per 100 pregnancies, compared with a local background rate of 12.8 (95% CI, 12.8-12.9) per 100 pregnancies. CONCLUSIONS: The passive surveillance provided rapid initial assessment of adverse events after 2009 H1N1 vaccination among pregnant women. Its findings were reassuring for the safety of 2009 H1N1 vaccines in pregnancy.