Towards early detection of adverse drug reactions: combining pre-clinical drug structures and post-market safety reports

A Drug Similarity-Based Bayesian Method for Early Adverse Drug Event Detection

INTRODUCTION

Biochemical drug similarity-based methods demonstrate successes in predicting adverse drug events (ADEs) in preclinical settings and enhancing signals of ADEs in real-world data mining. Despite these successes, drug similarity-based ADE detection shall be expanded with false-positive control and evaluated under a time-to-detection setting.

METHODS

We tested a drug similarity-based Bayesian method for early ADE detection with false-positive control. Under the tested method, prior distribution of ADE probability of a less frequent drug could be derived from frequent drugs with a high biochemical similarity, and posterior probability of null hypothesis could be used for signal detection and false-positive control. We evaluated the tested and reference methods by mining relatively newer drugs in real-world data (e.g., the US Food and Drug Administration (FDA)'s Adverse Event Reporting System (FAERS) data) and conducting a simulation study.

RESULTS

In FAERS analysis, the times to achieve a same probability of detection for drug-labeled ADEs following initial drug reporting were 5 years and ≥ 7 years for the tested method and reference methods, respectively. Additionally, the tested method compared with reference methods had higher AUC values (0.57-0.79 vs. 0.32-0.71), especially within 3 years following initial drug reporting. In a simulation study, the tested method demonstrated proper false-positive control, and had higher probabilities of detection (0.31-0.60 vs. 0.11-0.41) and AUC values (0.88-0.95 vs. 0.69-0.86) compared with reference methods. Additionally, we identified different types of drug similarities had a comparable performance in high-throughput ADE mining.

CONCLUSION

The drug similarity-based Bayesian ADE detection method might be able to accelerate ADE detection while controlling the false-positive rate.

Congenital anomalies associated with the use of cardiovascular drugs during pregnancy: a large-scale data analysis from the FAERS database

BACKGROUND

Cardiovascular drugs can cross the placenta during pregnancy, potentially exposing the fetus to teratogenic effects. However, ethical constraints on clinical trials with pregnant women limit safety data and result in inadequate drug labeling.

RESEARCH DESIGN AND METHODS

Using the FAERS database (2004-2023), we conducted a retrospective pharmacovigilance study analyzing adverse event reports involving congenital anomalies in newborns (<28 days). Signal detection methods included Reporting Odds Ratio (ROR), Proportional Reporting Ratio (PRR), Bayesian Confidence Propagation Neural Network (BCPNN), and Multi-item Gamma Poisson Shrinker (MGPS). Our analysis concentrated on the systems or organs involved in the signals, particularly those with higher report counts or signal values, to explore the association between drugs and congenital abnormalities.

RESULTS

Among 6,208 cases of congenital anomalies in newborns, 387 were linked to cardiovascular drugs, generating 97 signals for 16 drugs. Strong signals included sartans (renal failure, skeletal deformity), metoprolol (hypospadias, large-for-dates baby), amlodipine (gastrointestinal malformations), and statins, furosemide, and spironolactone (dysmorphism).

CONCLUSIONS

Enhanced monitoring is recommended for fetal malformations in women exposed to these drugs before or during pregnancy. While our findings suggest associations, they do not establish causality, highlighting the need for further research to ensure medication safety during pregnancy.

A Comparative Study Assessing the Incidence and Degree of Hyperkalemia in Patients on Unfractionated Heparin versus Low-Molecular Weight Heparin

Background

Heparin and its derivates, including unfractionated heparin (UFH) and low molecular weight heparin (LMWH), are among the most commonly used anticoagulants. Nonetheless, their use has been associated with hyperkalemia.

Objective

To determine and compare the incidence, magnitude, and potential risk factors of hyperkalemia in patients receiving UFH versus LMWH in a real-world clinical setting.

Methods

A retrospective observational study was conducted involving all adult hospitalized patients who received UFH, dalteparin or enoxaparin. Electronic medical records were reviewed over a 12-month period, collecting data on demographic, laboratory, comorbidity, and medication-related variables. Data were analyzed using multivariate logistic regression.

Results

A total of 929 patients met the eligibility criteria, with a mean age of over 40 years across all groups. Of these, 56.3%, 17.2%, and 15.7% experienced hyperkalemia with UFH, dalteparin and enoxaparin, respectively. The incidence of hyperkalemia was significantly higher with UFH compared to enoxaparin and dalteparin (p<0.001). Diabetes mellitus was associated with a higher incidence of hyperkalemia (OR 1.79, 95% CI 1.241-2.581, p=0.002), as was the concomitant use of co-trimoxazole (OR 2.244, 95% CI 1.137-4.426, p=0.02). Whilst chronic kidney disease and the use of two or more hyperkalemia-inducing agents were not statistically significant, they were retained in the model as they were associated with more than a 10% increase in the odds of hyperkalemia.

Conclusion

Heparin (UFH, LMWH) administration was associated with a risk of hyperkalemia particularly in patients with diabetes mellitus and those concurrently receiving co-trimoxazole.

Using multiple drug similarity networks to promote adverse drug event detection

The occurrence of an adverse drug event (ADE) has become a serious social concern of public health. Early detection of ADEs can lower the risk of drug safety as well as the expense of the drug. While post-market spontaneous reports of ADEs remain a cornerstone of pharmacovigilance, most existing signal detection algorithms rely on substantial accumulated data, limiting their applicability to early ADE detection when reports are scarce. To address this issue, we propose a label propagation model for generating enhanced drug safety signals using multiple drug features. We first construct multiple drug similarity networks using a range of drug features. We then calculate initial drug safety signals using conventional signal detection algorithms. These original signals are subsequently propagated across each drug similarity network to obtain enhanced drug safety signals. We evaluate our proposed model using two common signal detection algorithms on data from the FDA Adverse Event Reporting System (FAERS). Results demonstrate that enhanced drug safety signals with pre-clinical information outperform the standard safety signal detection algorithms on early ADE detection. In addition, we systematically evaluate the performance of different drug similarities against different types of ADEs. Furthermore, we have developed a web interface (http://drug-drug-sim.aimedlab.net/) to display our multiple drug similarity scores, facilitating access to this valuable resource for drug safety monitoring.

Accelerating adverse pregnancy outcomes research amidst rising medication use: parallel retrospective cohort analyses for signal prioritization

BACKGROUND

Pregnant women are significantly underrepresented in clinical trials, yet most of them take medication during pregnancy despite the limited safety data. The objective of this study was to characterize medication use during pregnancy and apply propensity score matching method at scale on patient records to accelerate and prioritize the drug effect signal detection associated with the risk of preterm birth and other adverse pregnancy outcomes.

METHODS

This was a retrospective study on continuously enrolled women who delivered live births between 2013/01/01 and 2022/12/31 (n = 365,075) at Providence St. Joseph Health. Our exposures of interest were all outpatient medications prescribed during pregnancy. We limited our analyses to medication that met the minimal sample size (n = 600). The primary outcome of interest was preterm birth. Secondary outcomes of interest were small for gestational age and low birth weight. We used propensity score matching at scale to evaluate the risk of these adverse pregnancy outcomes associated with drug exposure after adjusting for demographics, pregnancy characteristics, and comorbidities.

RESULTS

The total medication prescription rate increased from 58.5 to 75.3% (P < 0.0001) from 2013 to 2022. The prevalence rate of preterm birth was 7.7%. One hundred seventy-five out of 1329 prenatally prescribed outpatient medications met the minimum sample size. We identified 58 medications statistically significantly associated with the risk of preterm birth (P ≤ 0.1; decreased: 12, increased: 46).

CONCLUSIONS

Most pregnant women are prescribed medication during pregnancy. This highlights the need to utilize existing real-world data to enhance our knowledge of the safety of medications in pregnancy. We narrowed down from 1329 to 58 medications that showed statistically significant association with the risk of preterm birth even after addressing numerous covariates through propensity score matching. This data-driven approach demonstrated that multiple testable hypotheses in pregnancy pharmacology can be prioritized at scale and lays the foundation for application in other pregnancy outcomes.

Drug-induced liver injury associated with atypical generation antipsychotics from the FDA Adverse Event Reporting System (FAERS)

BACKGROUND

Recent studies have shown that liver enzyme abnormalities were not only seen with typical antipsychotics (APs) but also with atypical antipsychotics (AAPs). During the last 20 years, the hepatotoxicity of various antipsychotics received much attention. However, systematic evaluations of hepatotoxicity associated with APs are limited.

METHODS

All drug related hepatic disorders cases were retrieved from the FDA Adverse Event Reporting System (FAERS) database using standardized MedDRA queries (SMQ) from the first quarter of 2017 to the first quarter of 2022. Patient characteristics and prognosis were assessed. In this study, a case/non-case approach was used to calculate reporting odds ratio (RORs) and 95% confidence intervals (CIs). We calculated the drug-induced liver injury (DILI) RORs for each AAPs.

RESULTS

A total of 408 DILI cases were attributed to AAPs during the study period. 18.6% of these were designated as serious adverse event (SAE), which include death (19.74%), hospitalization (68.42%), disability (2.63%), and life-threatening (9.21%) outcomes. The RORs values in descending order were: quetiapine (ROR = 0.782), clozapine (ROR = 0.665), aripiprazole (ROR = 0.507), amisulpride (ROR = 0.308), paliperidone (ROR = 0.212), risperidone (ROR = 0.198), ziprasidone (0.131).

CONCLUSION

The result found in our study was that all AAPs didn't have a significant correlation with increased hepatotoxicity. Future analysis of the FAERS database in conjunction with other data sources will be essential for continuous monitoring of DILI.

Network-Based Methods and Their Applications in Drug Discovery

Drug discovery is time-consuming, expensive, and predominantly follows the "one drug → one target → one disease" paradigm. With the rapid development of systems biology and network pharmacology, a novel drug discovery paradigm, "multidrug → multitarget → multidisease", has emerged. This new holistic paradigm of drug discovery aligns well with the essence of networks, leading to the emergence of network-based methods in the field of drug discovery. In this Perspective, we initially introduce the concept and data sources of networks and highlight classical methodologies employed in network-based methods. Subsequently, we focus on the practical applications of network-based methods across various areas of drug discovery, such as target prediction, virtual screening, prediction of drug therapeutic effects or adverse drug events, and elucidation of molecular mechanisms. In addition, we provide representative web servers for researchers to use network-based methods in specific applications. Finally, we discuss several challenges of network-based methods and the directions for future development. In a word, network-based methods could serve as powerful tools to accelerate drug discovery.

An Early Adverse Drug Event Detection Approach with False Discovery Rate Control

Adverse drug event (ADE) is a significant challenge in clinical practice. Many ADEs have not been identified timely after the approval of the corresponding drugs. Despite the use of drug similarity network demonstrates early success on improving ADE detection, false discovery rate (FDR) control remains unclear in its application. Additionally, performance of early ADE detection has not been explicitly investigated under the time-to-event framework. In this manuscript, we propose to use the drug similarity based posterior probability of null hypothesis for early ADE detection. The proposed approach is also able to control FDR for monitoring a large number of ADEs of multiple drugs. The proposed approach outperforms existing approaches on mining labeled ADEs in the US FDA's Adverse Event Reporting System (FAERS) data, especially in the first few years after the drug initial reporting time. Additionally, the proposed approach is able to identify more labeled ADEs and has significantly lower time to ADE detection. In simulation study, the proposed approach demonstrates proper FDR control, as well as has better true positive rate and an excellent true negative rate. In our exemplified FAERS analysis, the proposed approach detects new ADE signals and identifies ADE signals in a timelier fashion than existing approach. In conclusion, the proposed approach is able to both reduce the time and improve the FDR control for ADE detection.

Application of an Integrative Drug Safety Model for Detection of Adverse Drug Events Associated With Inhibition of Glutathione Peroxidase 1 in Chronic Obstructive Pulmonary Disease

BACKGROUND

Chronic Obstructive Pulmonary Disease is characterised by declining lung function and a greater oxidative stress burden due to reduced activity of antioxidant enzymes such as Glutathione Peroxidase 1.

OBJECTIVES

The extent to which drugs may contribute to this compromised activity is largely unknown. An integrative drug safety model explores inhibition of Glutathione Peroxidase 1 by drugs and their association with chronic obstructive pulmonary disease adverse drug events.

METHODS

In silico molecular modelling approaches were utilised to predict the interactions that drugs have within the active site of Glutathione Peroxidase 1 in both human and bovine models. Similarities of chemical features between approved drugs and the known inhibitor tiopronin were also investigated. Subsequently the Food and Drug Administration Adverse Event System was searched to uncover adverse drug event signals associated with chronic obstructive pulmonary disease.

RESULTS

Statistical and molecular modelling analyses confirmed that the use of several registered drugs, including acetylsalicylic acid and atenolol may be associated with inhibition of Glutathione Peroxidase 1 and chronic obstructive pulmonary disease.

CONCLUSION

The integration of molecular modelling and pharmacoepidemological data has the potential to advance drug safety science. Ongoing review of medication use and further pharmacoepidemiological and biological analyses are warranted to ensure appropriate use is recommended.

iADRGSE: A Graph-Embedding and Self-Attention Encoding for Identifying Adverse Drug Reaction in the Earlier Phase of Drug Development

Adverse drug reactions (ADRs) are a major issue to be addressed by the pharmaceutical industry. Early and accurate detection of potential ADRs contributes to enhancing drug safety and reducing financial expenses. The majority of the approaches that have been employed to identify ADRs are limited to determining whether a drug exhibits an ADR, rather than identifying the exact type of ADR. By introducing the "multi-level feature-fusion deep-learning model", a new predictor, called iADRGSE, has been developed, which can be used to identify adverse drug reactions at the early stage of drug discovery. iADRGSE integrates a self-attentive module and a graph-network module that can extract one-dimensional sub-structure sequence information and two-dimensional chemical-structure graph information of drug molecules. As a demonstration, cross-validation and independent testing were performed with iADRGSE on a dataset of ADRs classified into 27 categories, based on SOC (system organ classification). In addition, experiments comparing iADRGSE with approaches such as NPF were conducted on the OMOP dataset, using the jackknife test method. Experiments show that iADRGSE was superior to existing state-of-the-art predictors.

Underreporting and Triggering Factors for Reporting ADRs of Two Ophthalmic Drugs: A Comparison between Spontaneous Reports and Active Pharmacovigilance Databases

(1) Aims of the study: calculating the underreporting ratio for two different medications, a fixed combination of 0.5% timolol + 0.2% brimonidine + 2.0% dorzolamide (antiglaucoma) and a fixed combination of sodium hyaluronate 0.1% + chondroitin sulfate 0.18% (artificial tears) for characterizing the features influencing the reporting of adverse drug reactions (ADRs) in spontaneous reporting. (2) Methods: The underreporting ratio was calculated by comparing the adverse drug reactions reported in the spontaneous reporting database for every 10,000 defined daily doses marketed and the adverse drug reactions from an active surveillance study for every 10,000 defined daily doses used for different drugs (antiglaucoma and artificial tears). The factors related to the report in spontaneous reporting through statistical tests were also determined. (3) Results: The underreporting ratio of spontaneous reporting was 0.006029% for antiglaucoma and 0.003552% for artificial tears. Additionally, statistically significant differences were found for severity, unexpected adverse drug reactions, and incidence of adverse drug reactions in females when compared with spontaneous reporting and active surveillance. (4) Conclusions: The underreporting ratio of ADRs related to ophthalmic medications indicates worry since the cornerstone of pharmacovigilance focuses on spontaneous reporting. Additionally, since underreporting seems to b selective, the role of certain aspects, such as gender, seriousness, severity, and unexpected ADRs, must be considered in future research.

Combining a Pharmacological Network Model with a Bayesian Signal Detection Algorithm to Improve the Detection of Adverse Drug Events

Introduction: Improving adverse drug event (ADE) detection is important for post-marketing drug safety surveillance. Existing statistical approaches can be further optimized owing to their high efficiency and low cost.

Objective: The objective of this study was to evaluate the proposed approach for use in pharmacovigilance, the early detection of potential ADEs, and the improvement of drug safety.

Methods: We developed a novel integrated approach, the Bayesian signal detection algorithm, based on the pharmacological network model (IC_PNM) using the FDA Adverse Event Reporting System (FAERS) data published from 2004 to 2009 and from 2014 to 2019Q2, PubChem, and DrugBank database. First, we used a pharmacological network model to generate the probabilities for drug-ADE associations, which comprised the proper prior information component (IC). We then defined the probability of the propensity score adjustment based on a logistic regression model to control for the confounding bias. Finally, we chose the Side Effect Resource (SIDER) and the Observational Medical Outcomes Partnership (OMOP) data to evaluate the detection performance and robustness of the IC_PNM compared with the statistical approaches [disproportionality analysis (DPA)] by using the area under the receiver operator characteristics curve (AUC) and Youden’s index.

Results: Of the statistical approaches implemented, the IC_PNM showed the best performance (AUC, 0.8291; Youden’s index, 0.5836). Meanwhile, the AUCs of the IC, EBGM, ROR, and PRR were 0.7343, 0.7231, 0.6828, and 0.6721, respectively.

Conclusion: The proposed IC_PNM combined the strengths of the pharmacological network model and the Bayesian signal detection algorithm and performed better in detecting true drug-ADE associations. It also detected newer ADE signals than a DPA and may be complementary to the existing statistical approaches.

Analysis of the Frequency and Onset Time of Hyponatremia/Syndrome of Inappropriate Antidiuretic Hormone Induced by Antidepressants or Antipsychotics

BACKGROUND

Hyponatremia and syndrome of inappropriate antidiuretic hormone (SIADH) is a potentially fatal adverse effect of antidepressants (ADs) and antipsychotics (APs), although its frequency and onset time have not been well documented.

OBJECTIVE

To analyze the frequency and onset time of AD- or AP-induced hyponatremia/SIADH.

METHODS

We used plural data-mining techniques to search the US Food and Drug Administration Adverse Event Reporting System (FAERS) database for reports on hyponatremia/SIADH induced by psychotropic drugs from January 2004 to June 2020. For each item, we assessed the reporting odds ratio, 95% CI, median onset time, and Weibull distribution parameters.

RESULTS

We identified 36 422 reports related to hyponatremia/SIADH. Signals were detected for all psychotropic drugs that we analyzed, except for clozapine. The median onset time of total AD-induced hyponatremia/SIADH was shorter than that of AP. For all ADs and APs except clozapine, hazards were considered to be the early failure type. In contrast, the hazard of clozapine was considered to be the random failure type. The limitations of this study included several reporting biases and the presence of confounding variables, particularly age.

CONCLUSION AND RELEVANCE

Most ADs and APs were found to be associated with a risk for hyponatremia/SIADH. In addition, sufficient attention should be paid to signs of hyponatremia/SIADH in the early phase when most ADs and APs are administered. These data are potentially useful for determining AD- or AP-induced hyponatremia/SIADH in the early stage and for preventing its further aggravation into a serious condition.

Nephrotoxicity of Herbal Products in Europe-A Review of an Underestimated Problem

Currently in Europe, despite the many advances in production technology of synthetic drugs, the interest in natural herbal medicines continues to increase. One of the reasons for their popular use is the assumption that natural equals safe. However, herbal medicines contain pharmacologically active ingredients, some of which have been associated with adverse effects. Kidneys are particularly susceptible to injury induced by toxins, including poisonous constituents from medicinal plants. The most recognized herb-induced kidney injury is aristolochic acid nephropathy connected with misuse of certain Traditional Chinese herbal medicines. Data concerning nephrotoxicity of plant species of European origin are scarce. Here, we critically review significant data of the nephrotoxicity of several plants used in European phytotherapy, including Artemisia herba-alba, Glycyrrhiza glabra, Euphorbia paralias, and Aloe). Causative mechanisms and factors predisposing to intoxications from the use of herbs are discussed. The basic intention of this review is to improve pharmacovigilance of herbal medicine, especially in patients with chronic kidney diseases.

Drug-induced liver injury associated with antiseizure medications from the FDA Adverse Event Reporting System (FAERS)

PURPOSE

Treatment with antiseizure medications (ASMs) confers a risk of drug-induced liver injury (DILI), especially for older ASMs. We sought to quantify recent reports of DILI attributed to both older and newer generation ASMs and survey newly marketed ASMs for hepatotoxicity in a large post-marketing database.

METHODS

We queried over 2.6 million adverse event reports made to the FDA Adverse Event Reporting System (FAERS) database between July 1, 2018 and March 31, 2020 for DILI due to ASMs commonly used in clinical practice. Patient characteristics and outcomes were assessed. We calculated the reporting odds ratio (ROR) of DILI for each individual ASM versus all non-ASM reports.

RESULTS

A total of 2175 DILI cases were attributed to an ASM during the study period. 97.2% of these were designated as serious reactions, which include death, hospitalization, disability, and other life-threatening outcomes. A number of older and newer generation ASMs were associated with DILI, specifically: carbamazepine (ROR 2.92), phenobarbital (ROR 2.91), oxcarbazepine (ROR 2.58), phenytoin (ROR 2.40), valproate (ROR 2.22), lamotrigine (ROR 2.06), clobazam (ROR 1.67), levetiracetam (ROR 1.56), and diazepam (ROR 1.53). However, increased odds of DILI were not seen with zonisamide, perampanel, stiripentol, lacosamide, clonazepam, pregabalin, felbamate, eslicarbazepine, cannabidiol, topiramate, gabapentin, ethosuximide, brivaracetam, or primidone. Vigabatrin, tiagabine, and rufinamide all had zero reports of DILI.

CONCLUSIONS

The majority of newer generation ASMs were not significantly associated with DILI. Future studies utilizing FAERS in conjunction with other data sources will be critical for the ongoing surveillance of DILI, particularly as newly marketed ASMs continue to enter into widespread clinical use.

Guillain-Barré syndrome in patients treated with immune checkpoint inhibitors

OBJECTIVE

Guillain-Barré syndrome (GBS) induced by immune checkpoint inhibitors (ICIs) has been occasionally reported in randomized clinical trials (RCTs), but the post-marketing data are quite limited. This study aimed to comprehensively examine GBS events secondary to ICI treatments in the real-world patients based on the Food and Drug Administration Adverse Event Reporting System (FAERS).

METHODS

Reports from January 2004 to March 2020 were extracted from the FAERS. GBS cases related to ICIs were identified to characterize their clinical features. The disproportionality and Bayesian analysis were performed for the detection of GBS signals associated with ICIs.

RESULTS

In total, 149 GBS reports with ICIs as suspect drugs were screened out. These events were found to be more prevalent in adults ≥ 45 years (63.09%) and males (63.09%). The onsets of GBS were variable with a median time of 38 (range 0-628) days after ICI initiation. The outcomes tended to be severe with 61.74% hospitalization and 22.82% death. GBS events were most commonly reported in ipilimumab plus nivolumab treatment (24.83%), and this combination therapy also yielded stronger signal for GBS than other therapies based on the highest reporting odds ratio (ROR = 12.43, two-sided 95% CI = 8.62, 17.93), proportional reporting ratio (PRR = 12.39, χ² = 300.90), information component (IC = 3.62, IC025 = 2.51) and empirical Bayes geometric mean (EBGM = 12.28, EBGM05 = 9.04).

CONCLUSION

As complements to the safety data from RCTs, the current pharmacovigilance research helps establish a more detailed overview of ICI-related GBS, which facilitates the understanding of this rare adverse drug effect.

Genetic Variants as Sudden-Death Risk Markers in Inherited Arrhythmogenic Syndromes: Personalized Genetic Interpretation

Inherited arrhythmogenic syndromes are the primary cause of unexpected lethal cardiac episodes in young people. It is possible that the first sign of the condition may be sudden death. Inherited arrhythmogenic syndromes are caused by genetic defects that may be analyzed using different technical approaches. A genetic alteration may be used as a marker of risk for families who carry the genetic alterations. Therefore, the early identification of the responsible genetic defect may help the adoption of preventive therapeutic measures focused on reducing the risk of lethal arrhythmias. Here, we describe the use of massive sequencing technologies and the interpretation of genetic analyses in inherited arrhythmogenic syndromes.