International Society of Pharmacovigilance (ISoP) 18th Annual Meeting, Geneva, November 2018

Title

Abstract 262: Passive Enhanced Safety Surveillance In Children Receiving Fluenz® Tetra Vaccination In England During The Early 2017-2018 Influenza Season.

Hazell1, S. Shakir* 1, J. Dentith2, E. Hammond3, D. Brooks3

1Drug Safety Research Unit, Southampton, 2AstraZeneca, Luton, United Kingdom, 3AstraZeneca, Gaithersburg, United States

Background

Fluenz® Tetra is a quadrivalent, live attenuated, intranasal, influenza vaccine recommended for use in children aged 2 to 17 years vaccinated as part of the seasonal influenza immunisation campaign in the UK.[1] Here, we present results from a third season of passive enhanced safety surveillance (ESS) for the vaccine in 2017-2018.

Objective

To measure and assess the frequencies of suspected adverse drug reactions (sADRs) in children receiving Fluenz® Tetra during the early 2017-2018 influenza season in England.

Methods

Vaccinees or parents/guardians received a Safety Report Card (SRC) to return if children experienced sADRs after vaccination with Fluenz® Tetra. At participating sites, 38 general practices and 36 primary schools in England, immunisation teams recorded numbers of SRCs distributed. The study was approved by an NHS Research Ethics Committee (Hampshire B South Central NRES Committee).

Results

Between 29th September 2017 and 2nd December 2017, 12,193 children were vaccinated at study sites, with 16 different vaccine lots used. In total, 10,793 SRCs were issued for children aged 2 to 17 including 5,189 children (48.1%) aged 2 to 4 years, 5,036 (46.7%) aged 5 to 10 years and 568 (5.3%) aged 11 to 17 years. Of 114 SRCs returned (by 9th January 2018 datalock), 101 reported at least one sADR (0.9% of all SRCs issued). The most frequently reported sADRs were rhinorrhoea (n=36), pyrexia (31) and cough (14). Three reported sADRs were classified as serious due to hospitalisation (0.03% of all SRCs issued) including nasal symptoms, general cold/flu symptoms and croup. The frequency of sADRs reported in the current ESS season was consistent with that reported in previous seasons.

Conclusion

Overall, the number of sADRs reported for Fluenz® Tetra in the 2017-2018 ESS remains low. Reported sADRs were typically minor expected events with serious sADRs rarely reported. The pattern and frequency of reported sADRs in the 2017-2018 season were consistent with previous seasons with no safety signals detected. Despite the limited number of reports received, the ESS method continues to be a useful extension to routine pharmacovigilance activities in the monitoring of possible sADRs.

Project sponsored by DSRU and AstraZeneca.

Reference

1. AstraZeneca. Fluenz® Tetra. Summary of Product Characteristics. Last updated: 22nd March 2018. Available at https://www.medicines.org.uk/emc/product/3296/smpc. Accessed: 9th May 2018

Title

Abstract 263: Utilisation and Safety of Asenapine in Primary Care in England: Results from a PASS

Sandeep Dhanda, Flavien Coukan, Lesley Wise, Saad Shakir

Background

Asenapine is approved in the EU for the treatment of manic episodes associated with bipolar I disorder in adults. A PASS was carried out to monitor the use and safety of asenapine using real-world primary care data in England; targeted safety outcomes were oral hypoaesthesia, oropharyngeal swelling, somnolence and sedation, acute allergic reactions and weight gain (defined as weight (kg) and/or BMI (kg/m2) increase of ≥7% at end of study vs. index).

Objectives

To describe the utilisation and safety of asenapine in primary care in England.

Methods

A cohort study identified patients from dispensed prescriptions of asenapine in England from January 2012-June 2016. Patient characteristics, drug utilisation and outcome data were collected from prescribing general practitioners via questionnaires sent at ≥3, and ≥12 months after the 1st prescription issued for each patient. Summary descriptive statistics were calculated.

Results

The cohort consisted of 122 patients with 3-month data of whom 56 patients had corresponding 12-month data (median age 44 years [IQR 35-55]; 66.4% female). The most frequent indication was bipolar disorder (n=81, 66.4%); for 10 patients (12.3%), the indication was specifically reported as bipolar I disorder. There were also a number of off-label indications e.g., schizophrenia (n=21, 17.2%). Asenapine was prescribed in accordance with the product label for 61 patients (50.0%) who were started on 5mg bd; 19 patients (15.6%) were prescribed 10mg bd.

For the 3-month cohort, incidence of targeted events was; oral hypoaesthesia n=0 (0.0%), oropharyngeal swelling n=1 (0.8%), sedation/somnolence n=5 (4.1%), acute allergic reactions n=0 (0.0%).

For the 12-month cohort, incidence of targeted events was; oral hypoaesthesia n=0 (0.0%), oropharyngeal swelling n=1 (1.8%), sedation/somnolence n=6 (10.7%), acute allergic reactions n=0 (0.0%), weight gain n=10 (17.9%).

In addition to the targeted outcomes other events of interest included four deaths, one of which was a completed suicide, and a case of foetal spina bifida diagnosed in a pregnant patient taking multiple psychiatric medications.

Conclusions

Asenapine was largely prescribed in accordance with prescribing recommendations. Event counts were small and no new safety signals were identified. This study design allowed for the timely collection of drug utilisation and safety data directly from prescribing GPs. However, as a result of the small cohort size, any conclusions from this study should be put into context with results from other post-marketing studies.

Title

Abstract 264: Utilisation and safety of asenapine in secondary care in the UK: Results from the Observational Safety Evaluation of Asenapine (OBSERVA) study

Sandeep Dhanda, Jan Slade, Vicki Osborne, Lesley Wise, Saad Shakir

Background

The OBSERVA specialist cohort event monitoring (SCEM) study, conducted in secondary care in the UK, formed part of an EU risk management plan to monitor the use and short-term safety of asenapine, licensed for the treatment of moderate to severe manic episodes associated with bipolar I disorder in adults. Targeted safety outcomes were oral hypoaesthesia, oropharyngeal swelling, somnolence and sedation, acute allergic reactions and weight gain (defined as weight (kg) and/or BMI (kg/m2) increase of ≥7% at end of study vs. index).

Objectives

To describe the utilisation and safety of asenapine prescribed in secondary care in the UK.

Methods

An observational cohort study identified patients via a network of psychiatrists in collaboration with the Mental Health Research Network from November 2012-September 2016. Patient characteristics, drug utilisation and outcome data were collected via questionnaires completed by specialists sent at baseline and after 12 weeks of observation. Summary descriptive statistics were calculated.

Results

The final cohort consisted of 125 patients; the median age was 44 years [IQR 35-52]; 63 (50.4%) were male. Asenapine was most frequently prescribed for bipolar disorder (n=81, 64.8%); 15 of these patients (18.5%) had bipolar I disorder reported. There were also a number of off-label indications e.g., schizophrenia (n=18, 14.4%). Asenapine was largely prescribed in accordance with the product label; 56 patients (44.8%) were started on 5mg bd and at 12 weeks 34 patients (27.2%) were taking 10mg bd.

The incidence of targeted events was as follows; oral hypoaesthesia n=9 (7.2%), oropharyngeal swelling n=0 (0.0%), sedation n=13 (10.4%), somnolence n=11 (8.8%), acute allergic reactions n=0 (0.0%), weight gain n=1 (0.8%).

No deaths or pregnancies were reported in the OBSERVA study.

Conclusions

Asenapine was largely prescribed in accordance with prescribing recommendations; the majority of patients were taking asenapine for bipolar disorder at the recommended dose. Event counts were small and no new safety signals were identified. This study provides valuable real-word data to support the post marketing risk-benefit profile of asenapine, however as a result of the small cohort size, any conclusions from this study should be put into context with results from other research evidence.

Title

Abstract 265: Risk Of Major And Clinically Relevant Non-Major (CRNM) Bleeding In Patients Prescribed Rivaroxaban In Primary Care In England

Sandeep Dhanda1,2, Miranda Davies1,2, Debabrata Roy1,2, Lesley Wise1, Saad Shakir1,2

Affiliations: 1 Drug Safety Research Unit, 2 University of Portsmouth

Background

Clinical trials and observational studies have reported bleeding risk in patients taking rivaroxaban. A PASS was carried out as part of the Risk Management Plan to monitor the safety and use of rivaroxaban using real-world primary care data in England.

Objectives

To estimate the risk of Major and CRNM bleeding in patients prescribed rivaroxaban for Stroke prevention in non-valvular AF (SPAF) and for the prevention and/or treatment of Deep Vein Thrombosis and/or Pulmonary Embolism DVT/PE in primary care.

Methods

Patients identified from dispensed prescriptions in England (2012-2016). Detailed questionnaires sent to general practitioners (GPs) at ≥3 and ≥12 months of observation collected information on risk factors for bleeding (HAS-BLED) and bleeding outcomes. Summary descriptive statistics and 12-month risk estimates were calculated.

Results

Cohort = 17546 patients: 10225 patients with AF (58.3% of cohort, median age 78 years [IQR 70-84], 5253 (51.4%) male); 5959 patients with DVT/PE (34.0 % of cohort, median age 66 years [IQR 50-78]; 3197 (53.6%) female). In both groups, the median HAS-BLED score was 1 (IQR 1-2, 0-1, respectively) reflecting a low risk of major bleeding.

AF group: Risk Major + CRNM bleeding 8.3% ([95% CI 7.8, 8.9]; n=825). Risk Major bleed 2.4% ([95% CI 2.1, 2.7]; n=239), CRNM bleeding 6.0% ([95% CI 5.5, 6.4]; n=592). Major bleeding further stratified by site: gastrointestinal (GI) (1.2%; n=117), urogenital (UG) (0.1%; n=13), intracranial (IC) (0.4%; n=42), all other critical organ (excluding IC) (0.3%; n=26) and all non-critical organ sites (0.4%; n=44).

DVT/PE group: Risk Major + CRNM bleeding 4.2% ([95% CI 3.7, 4.7]; n=240). Risk Major bleed 1.4% ([95% CI 1.1, 1.7]; n=82), CRNM bleeding 2.8% ([95% CI 2.4, 3.3]; n=162). Major bleeding further stratified by site: GI (0.7%; n=38), UG (0.3%; n=18), IC (0.2%; n=12), all other critical organ (excluding IC) (0.1%; n=4) and all non-critical organ sites (0.2%; n=10).

Conclusions

For the primary outcome of major bleeding, the estimates of risk in the AF and DVT/PE rivaroxaban user populations were overall low and consistent with those estimated from clinical trial data. Differences in methodologies and analysed study populations prevent meaningful comparisons with other studies. This study design has unique strengths, including the collection of timely, granular data directly from prescribing GPs, however selective reporting of outcomes and selection bias might be present, and should be considered when interpreting results.

Title

Abstract 266: Multi-Level Modelling To Investigate Factors Impacting Prescribing Variability

Roy* 1, 2, L. Wise1, S. Shakir1, 2

1Drug Safety Research Unit, Southampton, 2University of Portsmouth, Portsmouth, United Kingdom

Background

Prescribing guidelines influence treatment choice based on patient and healthcare system factors. Multi-level modelling can provide insight into sources of variability in healthcare, especially where nested hierarchical structures exist.[1] A Specialist Cohort Event Monitoring study investigated the safety and use of rivaroxaban in clinical use, with a warfarin cohort for context.

Objective

Study to investigate prescribing variability using Multi-level modelling.

Methods

Data on NHS acute trusts in England/Wales (e.g. population size, trust type) were linked to study patient demographic and drug utilization data, and prescriber details (e.g. degree, specialty).

Using Multi-level modelling we explored the influence of patient, prescriber and trust characteristics on prescribing variability in 2106 rivaroxaban (59%) vs. 1468 warfarin (41%) adult patients nested in 780 prescribers, nested in 73 trusts. The majority of patients had an indication of DVT/PE (56.4%) or non-valvular AF (AF) (41.2%). The binary outcome was rivaroxaban or warfarin treatment.

Variance components estimate the variability accounted for by each level in the model and are expressed as: Median Odd Ratios (MOR – median relative increase in odds of rivaroxaban treatment if patient changed prescriber (PR) or trust (T)); Proportional Change in Variance (PCV) between models when successively adding fixed effects.

Results

DVT/PE group:

Adjusting for patient factors, MORT=6.9 (PCV=-0.6%); MORPR=2.8 (PCV=3.8%).

Adjusting for patient and prescriber factors, MORT=6.8 (PCV=-2.9%); MORPR=2.6 (PCV=-15.4%).

Adjusting for patient, prescriber and trust factors, MORT=4.9 (PCV=-30.2%); MORPR=2.6 (PCV=3.4%).

Differences between trusts and prescribers (in trusts) are important in treatment choice; trust being more influential. Some patient factors had a relatively large effect on odds of treatment choice although the absolute number of patient impacted was often small. Trust type, was shown to be associated with the odds of treatment choice [Final model: foundation vs acute trusts OR 4.0 (95%CI 1.5, 9.9)].

Data on AF and all indications (combined) will be included in the final presentation.

Conclusion

This study highlights the utility of Multi-level modelling in exploring patient and non-patient factors in nested hierarchical healthcare settings. Prescribing variability appears dominated by differences between trusts and prescribers (in trusts). Some patient factors were important in treatment choice, but PCV between models suggest that accounting for patient differences does not fully explain the variance between prescribers (in trusts) and between trusts.

References

1. Zuidgeest MG, van Dijk L, Spreeuwenberg P, Smit HA, Brunekreef B, Arets HG, et al. What drives prescribing of asthma medication to children? A multilevel population-based study. Ann Fam Med. 2009;7(1):32-40.

Title

Abstract 267: Incidence Of Major And Clinically-Relevant Non-Major Bleeding In Patients Prescribed Rivaroxaban: Results From The ROSE Study

A. Evans 1, 2, M. Davies 1, 2, L. Wise1, S. Shakir1, 2*

1Drug Safety Research Unit, Southampton, 2University of Portsmouth, Portsmouth, United Kingdom

Background

The ROSE (Rivaroxaban Observational Safety Evaluation) Specialist Cohort Event Monitoring (SCEM) study was conducted to monitor the safety and use of rivaroxaban for Stroke Prevention in patients (pts) with non-valvular AF (SPAF) and the prevention and/or treatment of Deep Vein Thrombosis and/or Pulmonary Embolism (DVT/PE) in the secondary care setting in England and Wales. The study complements another study conducted in primary care by focusing on the acute phase of treatment.

Objectives

To estimate the incidence of major and clinically-relevant non-major (CRNM) bleeding (ISTH definition) in patients prescribed rivaroxaban for the first time for SPAF and DVT/PE indications.

Methods

Patients were identified through speciality groups (2013-2016), supported by UK Clinical Research Networks. Risk factors for bleeding (HAS-BLED) and bleeding outcomes were collected via detailed questionnaires completed by hospital specialists at baseline and ≥ 12 weeks. Summary descriptive statistics and 12 week incidence risk and rate (per 100 patient years) were calculated.

Results

The cohort consisted of 2542 pts including 965 patients with AF (38.0% of cohort; median age 76 years [IQR 69, 83], 517 [53.6%] male), median HAS-BLED score 2 (IQR 1-3) (moderate risk of major bleeding) and 1532 patients with DVT/PE (60.3% of cohort; median age 63 years [IQR 48, 73], 836 [54.6%] male), median HAS-BLED score 1 (IQR 0-2) (low risk of major bleeding).

AF group: Rate: Major + CRNM bleeding 28.2 (95% CI 21.0, 37.1; n=51), Major Bleed 5.5 (95% CI 2.6-10.1; n=10), CRNM bleeding 22.7 (95% CI 16.3-30.8; n=41).

Major Bleed risk in pre-specified sites: gastrointestinal (GI) (0.2%; n=2), urogenital (UG) (0.2%; n=2), intracranial (IC) (0.2%; n=2) and all other critical organ (excluding IC) (0.1%; n=1).

DVT/PE group: Rate: Major + CRNM bleeding 36.2 (95% CI 29.4, 44.1; n=98), Major Bleed 8.3 (95% CI 5.3, 12.5; n=23), CRNM bleeding 27.6 (95% CI 21.7, 34.6; n=75).

Major Bleed risk in pre-specified sites: GI (0.7%; n=11), UG (0.3%; n=5), IC (0.1%; n=1) and all other critical organ (excluding IC) (0.0%; n=0).

Conclusions

The incidence of major bleeding was low and no new safety concerns were raised. Unique aspects of this study design enabled collection of highly detailed information from cardiologists, faciliating accurate calculation of risk scores and the application of clinical trial outcome definitions, in a real-word setting. Differences in methodology between this and other study designs prevents meaningful comparison.

Title

Abstract 268: Characteristics of Patients Started on a Reduced Dose of Rivaroxaban for Atrial Fibrillation – Results from the ROSE Study

Miranda Davies 1, 2, Alison Evans 1, 2, Flavien Coukan 1,2 , Lesley Wise1, Saad Shakir 1,2

Affiliations: 1 Drug Safety Research Unit, 2 University of Portsmouth

Background

Rivaroxaban is used to treat a range of conditions, including prevention of stroke and systemic embolism in patients with AF. This indication requires a fixed once-daily (od) dose (20mg) with dosage reduction to 15 mg once daily only recommended in patients (pts) with renal impairment (creatinine clearance 15-49 ml/min). However other clinical factors might be considered by the prescriber when choosing an initial dose.

Objectives

To describe the clinical characteristics of pts prescribed an initial dose of (≥ 20mg and <20 mg daily) for treatment of AF in secondary care.

Methods

The ROSE study was a specialist cohort event monitoring study of pts prescribed rivaroxaban. Specialists provided information from medical records via questionnaires at baseline and ≥ 12 weeks. Information on the prescribed initial dose and baseline characteristics including risk factors for bleeding (HAS-BLED criteria) was collected. For pts prescribed a reduced initial dose (<20 mg daily), we assessed clinical characteristics vs. pts prescribed ≥ 20mg daily.

Results

The cohort consisted of 965 pts with AF; of whom (18.3%, n=177), were prescribed an initial dose of <20 mg daily (15mg od [n=173]; 10mg od [n=3]; 4mg od [n=1]). The majority were female (57.6%, median age 84 yrs) vs. pts prescribed ≥ 20mg daily (20mg od [n=750]; 30mg od [n=16]), where the majority were male (56.4%, median age 75 yrs). 36/177 (20.3%) pts had a history of CKD 3-4 or 5. For 22 patients the start dose was unknown. Frequency of HAS-BLED criteria in pts prescribed <20 mg daily vs. ≥ 20mg daily was: Hypertension (37.3% vs. 38.5%), Abnormal renal function (Chronic dialysis, renal transplant, serum creatinine ≥ 2.3 mg/dL or 200 µmol/L) (3.4% vs. 1.2%), Abnormal liver function (0.6% vs. 1.0%), History of Stroke (31.1% vs. 30.4%), History of Bleeding/predisposition (18.6% vs. 16.3%), Labile INR (N/A), Age ≥65 yrs (96.1% vs. 81.2%), Drug therapy (52.0% vs. 51.4%), Alcohol (0% vs. 4.8%).

Conclusion

In pts prescribed a reduced total daily dose of rivaroxaban <20 mg od, there were a higher proportion of females and those aged ≥65 yrs. Patients prescribed this lower dose were also more likely to have abnormal renal function, defined as per HAS-BLED, than those patients prescribed ≥ 20mg daily. Although the label recommends dose reduction in AF pts with renal impairment, other factors seem to impact the choice of the initial dose in the clinical setting.

Title

Abstract 269: Assessment of Start Dose in Patients Prescribed Rivaroxaban for Atrial Fibrillation with Chronic Kidney Disease-Results from the ROSE Study

Miranda Davies 1, 2, Alison Evans 1,2, Flavien Coukan 1, 2, Lesley Wise1, Saad Shakir 1,2

Affiliations: 1 Drug Safety Research Unit, 2 University of Portsmouth

Background

Rivaroxaban is used to treat patients with non-valvular AF for the prevention of stroke and systemic embolism. This requires a fixed daily dose (20mg) with dosage reduction only recommended in patients with a reduced creatinine clearance (15-49 ml/min), a range encompassed within chronic kidney disease (CKD) stage 3-4 (eGFR 15-59) although CKD 3a (eGFR 45-59) is largely outside this range. Rivaroxaban is not recommended in patients with severe renal impairment (creatinine clearance < 15 ml/min).

Objectives

To assess starting dose amongst patients with chronic kidney disease stage 3-4 or 5.

Methods

The ROSE study was a specialist cohort event monitoring study of patients prescribed rivaroxaban. Specialists provided information via detailed questionnaires at baseline (and ≥ 12weeks). Baseline characteristics included starting dose and presence of either CKD stage 3-4 (eGFR 15-59) or CKD stage 5 (eGFR < 15). Starting dose was examined amongst patients with CKD stage 3-4 and 5 to assess how many patients had a reduced starting dose of less than 20mg od.

Results

The cohort consisted of 965 patients with AF: 75 patients with history of either stage 3-4 CKD (n=73, 7.6%) or stage 5 CKD (n=2, 0.2%). Of the patients with CKD stage 3-4, 36 (49.3 %) were started on <20 mg daily (15mg od [n=35]; 10mg od [n=1]). 35 patients (48.0 %) were started on ≥ 20mg (20mg od [n=34]; 30mg od [n=1]). Starting dose was missing for two patients (2.7%). Neither patient with CKD stage 5 had a dose reduction.

Conclusion

Our results suggest that amongst patients with CKD stage 3-4, approximately half were started on the recommended reduced dose of < 20mg od. Not all patients with CKD stage 3-4 would be recommended a dose reduction as per the product label. A UK audit suggests that approximately 20% of patients in the UK with CKD 3 are 3b rather than 3a although the frequency may be higher in a hospital cohort.

Title

Abstract 270: Distribution of CHA2DS2-VASc Scores in Patients with Atrial Fibrillation Treated with Rivaroxaban in Primary vs. Secondary Care Settings

Miranda Davies1,2, Lesley Wise1, Saad Shakir1,2

Affiliations: 1 Drug Safety Research Unit 2 University of Portsmouth

Background

Since its incorporation in the European Society of Cardiology guidelines in 2010, the CHA2DS2-VASc score is widely used to characterise the risk of stroke in patients (pts) with atrial fibrillation (AF). It is frequently calculated in pharmacoepidemiological studies through retrospective application of the component criteria included in the risk score. CHA2DS2-VASc Scores of 0, 1, or ≥2 indicate low, moderate, or high stroke risk, respectively.

Objectives

To describe the distribution of CHA2DS2-VASc scores and individual stroke risk components in two cohorts of pts prescribed rivaroxaban for AF in primary (1 o) vs. secondary (2 o) care.

Methods

Two PASS were conducted to investigate the safety of rivaroxaban in pts with AF in 1o care (Modified-Prescription Event Monitoring) and 2 o care (Specialist Cohort Event Monitoring) (2012-2016). Baseline characteristics were provided by general practitioners (1o care) and specialist prescribers (2 o care) using customised questionnaires. An algorithm was used to compute a CHA2DS2-VASc score (0 – 9) for each pt from fixed response options or open questions, according to published guidelines.

Results

The response rate for baseline questionnaires was lower in 1o care vs. 2 o care (22.3% vs. 99.7%). The 1o care cohort consisted of 10225 pts with a primary indication of AF (median age 78 yrs [IQR 70-84]; 5253 [51.4%] male). The median CHA2DS2-VASc score was 4 (IQR 3-5) reflecting a high risk of stroke. The 2 o care cohort consisted of 965 pts with a primary indication of AF (median age 76 yrs [IQR 69-83]; 517 [53.6%] male), with a median CHA2DS2-VASc score of 4 (IQR 3-6). There were a higher proportion of pts with a score of 6 – 9 in 2 o care vs. 1 o care. The proportions of pts in 1o care vs. 2 o care with each criterion were:

• Age 65-74 yrs (25.7% vs. 25.8%)
• ≥75 yrs (61.8% vs. 58.0%)
• Female (48.6% vs. 46.5%)
• Congestive Heart Failure/Left ventricular Dysfunction (14.3% vs. 14.6%)
• History Hypertension (82.6% vs. 73.2%)
• History Stroke, Transient Ischaemic Attack (TIA) or Thromboembolism (TE) (19.8% vs. 46.9%)
• Vascular disease (11.3% vs. 26.9%)
• Diabetes Mellitus (17.2% vs. 18.8%)

Conclusion

These results suggest that pts initiated rivaroxaban for AF in 2 o care are more likely to have a history of stroke, TIA, TE or vascular disease than pts treated in 1o care. This may mean that pts considered to be at ‘higher risk’ with greater co-morbidities are more likely to be managed in 2 o care, or that co-morbidities may be less well recorded in 1o care.

Reference

Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137(2):263-72.

Title

Abstract 271: Impact analyses of EU pharmacovigilance interventions on public health burden

S. Lane* 1, E. Lynn1, 2, J. Slattery3, S. Shakir1, 2
Affiliations: 1 Drug Safety Research Unit 2 University of Portsmouth 3 European Medicines Agency UK

Background

Adverse drug reactions (ADRs) account for 5% hospital admissions and 197,000 deaths each year, amounting to a societal cost of €79billion.[1, 2] These ADRs have large impacts on public health in terms of morbidity and mortality. Since the implementation of the 2012 pharmacovigilance (PV) legislation, monitoring success of pharmacovigilance (PV) has become commonplace; however impact of PV regulatory interventions on public health remains mostly unquantified.[3] There is a current need for research identifying the best technique for measuring impact of these actions, and guidelines on reporting impact analyses.[3]

Objectives

Interim analysis to identify predictive modelling techniques for estimation of public health impact of PV regulatory interventions.

Methods

Prescription products used in primary care whose marketing authorisations were withdrawn, revoked or suspended in France, Germany or the UK between 2012 and 2016 were previously identified.[4] The French, German and British Health Improvement Network (THIN) databases were used to estimate usage of the identified products for the year prior to PV action. Systematic searches of the PubMed/MEDLINE database, and European Medicines Agency and national competent authority websites and documents were conducted to identify quantitative studies for the product and ADR of interest. Study data were meta-analysed to produce an overall weighted odds ratio for each ADR relating to each product. Background risk was estimated using THIN data. Modelling of usage figures, odds ratios and background risk allowed the public health impact of the intervention to be estimated in terms of morbidity attributable to the withdrawn, revoked or suspended product.

Results

18 products were considered for impact analysis; 9 were excluded as no quantitative studies were identified for the respective ADRs. This interim analysis on a subset of these products provides a prediction of the number of ADRs avoided as a result of the included products’ marketing authorisation withdrawal, revocation or suspension, and an estimation of the public health impact of these PV actions using changes in morbidity as an indicator.

Conclusion

This study identifies a method for predicting public health impact of PV interventions, based on drug utilisation data and expected changes in morbidity. This interim analysis tested the method on a subset of products. Further validation of the predictive modelling method will be undertaken before completion of the study.

References

1. Mammi M, Citraro R, Torcasio G, Cusato G, Palleria C, di Paola ED. Pharmacovigilance in pharmaceutical companies: An overview. J Pharmacol Pharmacother. 2013;4(Suppl 1):S33-7.
2. Mazzitello C, Esposito S, De Francesco AE, Capuano A, Russo E, De Sarro G. Pharmacovigilance in Italy: An overview. J Pharmacol Pharmacother. 2013;4(Suppl 1):S20-8.
3. Goedecke T, Morales D, Pacurariu A, Kurz X. Measuring the impact of medicines regulatory interventions – systematic review and methodological considerations. Br J Clin Pharmacol. 2017.
4. Lane S, Lynn E, Shakir S. Investigation assessing the publicly available evidence supporting postmarketing withdrawals, revocations and suspensions of marketing authorisations in the EU since 2012. BMJ Open. 2018;8(1).

Title

Abstract 272: Changes in Evidence Supporting Post-Marketing Withdrawal of Marketing Authorisation in the EU

S. Lane* 1, E. Lynn1, 2, S. Shakir1, 2
Affiliations: 1 Drug Safety Research Unit 2 University of Portsmouth

Introduction

Market withdrawal of a product affects patients, healthcare professionals, regulators and manufacturers, therefore these decisions should be based on the most robust evidence available.[1] EU pharmacovigilance (PV) legislation was updated in 2012 and led to modernisation of PV systems in Europe. It is necessary to explore whether evolution of PV leads to changes in evidence supporting PV actions (e.g. withdrawals), and how the use of supporting evidence has adapted over time.

Objectives

To evaluate changes in evidence supporting post-marketing product withdrawals from 1999-2016.

Methods

Using 3 previously published studies, it was possible to compare evidence used to support post-marketing PV actions in the EU, spanning the period 1999 to 2016.[1-3] The number of withdrawals per year was calculated. Study designs supporting post-marketing withdrawals in 1999-2016 were evaluated. Changes to evidence supporting product withdrawals were assessed.

Results

A total of 31 market withdrawals occurred in the EU between 1999 and 2016 and were included in these publications, with an average of 1.72 per year. The number of withdrawals increased between 1999 to its peak in 2010 (5 withdrawals in 2010), and decreased from 2011 to 2016. In general the number of different study designs used to justify withdrawals increased between 1999 and 2016. Spontaneous and published case reports were the most common evidence source, involved in 29 withdrawals (93.5%). This evidence alone led to 4 (12.9%) withdrawals. The least common sources of supporting evidence were epidemiological studies (case-control n=6; cohort n=7). Other sources included randomised clinical trials (n=20), animal studies (n=13) and meta-analyses (n=7). The average number of distinct evidence types used was 3.1 per withdrawal. Six (19.4%) withdrawals were justified by just one source of evidence, with the remaining 25 (80.6%) withdrawals supported by at least 2 different types of evidence.

Conclusion

The number of products withdrawn from the EU market appears to have increased until 2010, and has since decreased. Although there is an apparent increase in different types of evidence being used to justify post-marketing withdrawals, spontaneous and published case reports remain a vital source of evidence for regulators.

References

1. Clarke A, Deeks JJ, Shakir SA. An assessment of the publicly disseminated evidence of safety used in decisions to withdraw medicinal products from the UK and US markets. Drug Saf. 2006;29(2):175-81.
2. Lane S, Lynn E, Shakir SAW. Measuring the Impact of Product Withdrawals and Other Major Pharmacovigilance Actions on Public Health Burden in the EU: Methodological Considerations. Drug Saf; 2017 Oct 15-18; Liverpool (UK): International Society of Parmacovigilance; 2017. p. 937-1045.
3. McNaughton R, Huet G, Shakir S. An investigation into drug products withdrawn from the EU market between 2002 and 2011 for safety reasons and the evidence used to support the decision-making. BMJ Open. 2014;4(1):e004221.

Title

Abstract 273: A Cross-Sectional Study of Product and Batch Traceability for Biologics in Clinical Practice and Adverse Drug Reaction Reporting in the United Kingdom

K Klein 1, 2, 3, L Hazell 4, P Stolk 1, 2, 3, S Shakir 4

Affiliations: 1Lygature, Utrecht, The Netherlands. 2Utrecht University, Utrecht, The Netherlands. 3Exon Consultancy, Amsterdam, The Netherlands, 4Drug Safety Research Unit (DSRU), Southampton, United Kingdom

Background

Biological medical products (‘biologics’) are derived from living cells or organisms. Due to their complex molecular structures and sensitivity to changes in the manufacturing process, variations may exist between biologics with the same active substance from different manufactures (e.g. ‘biosimilars’) or between different batches of the same product.

It is of vital importance that biologics are traceable by brand name and batch number when reporting an adverse drug reaction (ADR) but such product details are not always provided [1]. In a recent study in the Netherlands, the brand name was identifiable in 76% of ADR reports submitted for biologics, whereas only 5% contained a batch number [2]. Shortcomings in the recording and tracing of product and batch information in clinical practice are believed to be associated with the limited traceability of biologics found in ADR reports in the Netherlands.

Objectives

The UK BIO-TRAC study was initiated to assess the current status of traceability of brand names and batch numbers for (recombinant) biologics in UK clinical practice within the hospital setting and in the UK ADR reporting database (Yellow Card Scheme).

Methods

An on-line survey has been set up for hospital pharmacists in England to capture information on recording practices that contribute to the traceability of brand names and batch numbers for biologics in UK clinical practice and to gather the views and experiences of hospital pharmacists on current bottlenecks and how to improve the current systems. In addition, a comprehensive analysis of 9036 ADR reports for biologics submitted to the UK Yellow Card Scheme between January 2009 and September 2017 will assess the extent to which brand name or batch number details are included in these reports. This includes a stratified analysis by reporter type, therapeutic product classes, calendar year and other variables.

Results

The results of the UK BIO-TRAC study will be presented in the context of both objectives, the assessment of the recording practices for product specific details in routine clinical practice and the ADR data analysis of the MHRA Yellow Card database.

References

1. Directive 2010/84/EU of the European Parliament and of the Council of 15 December 2010, amending as regards pharmacovigilance, Directive 2001/83/EC on the Community code relating to medicinal products for human use. Available from: http://ec.europa.eu/health/files/eudralex/vol-1/dir_2010_84/dir_2010_84_en.pdf
2. Klein K, Scholl JHG, Vermeer NS, Broekmans AW, Van Puijenbroek EP, De Bruin ML, et al. Traceability of Biologics in The Netherlands: An Analysis of Information-Recording Systems in Clinical Practice and Spontaneous ADR Reports. Drug Saf. 2016 Feb;39(2):185–92.

Disclaimer

The study is conducted by DSRU Education and Research Ltd in collaboration with Lygature and is funded by a grant from ABPI.

This study is based in part on data obtained under licence from the UK Medicines and Healthcare Products Regulatory Agency. However, the interpretation and conclusions contained in this study are those of the author/s alone.

Title

Abstract 274: Risk Factors of Major Bleeding in Patients Prescribed Rivaroxaban in Primary Care in England: Based on a Modified-Prescription Event Monitoring Study

Roy* 1, 2, S. Dhanda1, 2, L. Wise1, S. Shakir1, 2

1Drug Safety Research Unit, Southampton, 2University of Portsmouth, Portsmouth, United Kingdom

Background

Clinical trials and observational studies have reported bleeding risk in patients taking oral anticoagulants. It is valuable to understand the predictors for major bleeding in patients prescribed rivaroxaban in primary care in England.

Objectives

Multivariable logistic regression (MLR) analyses to explore potential risk factors for major bleeding within gastrointestinal (GI), urogenital (UG) and intracranial (IC) sites.

Methods

A case/non-case design was used to study the association between clinical risk factors and major bleeding in a cohort of rivaroxaban patients (N=17546) in a single-arm Modified-Prescription Event Monitoring study identifying patients from dispensed prescriptions in England (2012-2016), followed for 12 months. Clinical risk factors for bleeding and bleeding outcomes were collected from prescribing general practitioners via questionnaires sent at ≥3, and ≥12 months observation. Univariate and multivariable logistic regression analyses were performed to examine the association for each different site. Multivariable analyses models were based on 2 different clinical approaches; Clinical Risk Factors Selection (CS) model and HAS-BLED (HB) model. The CS model included all reported clinical risk factors for bleeding. The HAS-BLED model included the HAS-BLED clinical risk score categories (low, moderate or high risk) and gender. Statistically significant (p<0.05) associations from the MLR models are presented in the results section.

Results

Risk factors for Major GI bleed (n=176)

CS Model: Age 65-74 years vs <65 years: OR=2.4 [95% CI 1.3, 4.6]; Age ≥75years vs <65 years: OR=4.2 [95% CI 2.3, 7.5]; Predisposition to or history of bleeding: OR=4.8 [95% CI 3.1, 7.5] HB Model: Moderate vs low: OR=4.0 [95% CI 2.1, 7.6]; High vs low: OR=8.9 [95% CI 4.0, 19.9] Risk factors for Major UG bleed (n=36) CS Model: Age 65-74 years vs <65 years: OR=0.2 [95% CI 0.1, 0.7]; Females vs males: OR= 2.9 [95% CI 1.4, 6.1]; Malignancy: OR=2.6 [95% CI 1.1, 6.3] HB Model: Females vs males: OR=2.7 [95% CI 1.3, 5.6] Risk factors for Major IC bleed (n=57) CS Model: Age ≥75 years vs <65 years: OR=2.8 [95% CI 1.1, 6.9]; History of cerebrovascular accident/transient ischaemic attack (CVA/TIA) (including haemorrhagic CVA): OR=2.2 [95% CI 1.3, 3.9]; Predisposition to or history of bleeding: OR 2.6 [95% CI 1.0, 6.7] HB Model: Moderate vs low: OR=3.3 [95% CI 1.2, 9.1]; High vs low: OR=9.0 [95% CI 2.5, 31.9]

Conclusion

Use of the case/non-case design to explore risk factors for bleeding in rivaroxaban patients using the CS model identified age ≥65 and history of bleeding or predisposition as statistically significant prognostic factors for major GI bleeds. For major UG bleeds, age group 65-74 years, female gender (which may be related to vaginal bleeding including menorrhagia) and malignancy were statistically significant in the CS model. For major IC bleeds, age ≥75 years, history of CVA/TIA and history of bleeding or predisposition to bleeding were identified as statistically significant risk factors in the CS model. The HB model showed that moderate and high risk scores were statistically significant risk factors for major GI and IC bleeds and gender statistically significant for major UG bleeds. Overall, findings from the CS model and the HB model are in keeping with known clinical risk factors for bleeding.