A collaborative study of the impact of N-nitrosamines presence and ARB recall on ARB utilization – results from IQVIA™ Disease Analyzer Germany

Study design, setting and participants

A drug utilization study was conducted between January 2014 and June 2020 using descriptive statistics, supplemented by an interrupted time series (ITS) analysis to study changes in prescribing trends for individual ARBs before and after the first ARB recalls in July 2018. All patients with at least one ARB prescription during the study period were included in the study. Prescribing of individual ARBs over time was studied as the monthly proportion of all ARB use and as the quarterly proportion of incident ARB use, respectively.

Patients with at least two ARB prescriptions on separate dates during the study period with a minimum look-back period of 365 days prior to the first ARB prescription (referred to as the first ARB treatment episode) during the study period were followed up until they switched to another ARB or an ACE inhibitor, discontinued the ARB treatment, or were lost to follow-up (end of data). For this purpose, the first ARB prescription date during the study period was the index date for the patient. Patients prescribed more than one ARB (e.g. both valsartan and candesartan) on the index date (n = 246) were excluded. For each individual ARB, the proportion of switches was considered in relation to all patients with ARB use during the quarter. The pattern of ARBs prescribed after the switch (i.e. the ARB that the patient switched to) was also assessed. Finally, the rate of switching to an alternative ARB during person-time of follow-up was analysed before and after the ARB recalls.

Data source

The study was based on prescribing data from general practitioner (GP) practices in the IQVIA™ Disease Analyzer database in Germany (DA Germany).

The study used DA Germany version June 2020. DA Germany collects computerised information from specialised and GP practices throughout Germany since 1992. Around 3% of GP practices from the different regions in Germany are included in the database. Data from DA Germany have been shown to be reasonably representative of German healthcare statistics for demographics and certain diseases [11, 12].

DA Germany contains information about prescriptions issued by physicians, whereas dispensing information from pharmacies is not available. Prescribed medicines in DA Germany are coded using EphMRA ATC codes and generic substance names for the active ingredients.

Exposure variables

ARBs were identified using EphMRA ATC codes C09C (angiotensin II antagonists, plain) and C09D (angiotensin II antagonists, combinations). ACE inhibitors were identified using EphMRA ATC codes C09A (ACE inhibitors, plain) and C09B (ACE inhibitors, combinations). Information on the prescribed ARB was used. The pharmacy could dispense an alternate product, with the same ARB, strength and formulation. ARBs prescribed in DA Germany included azilsartan, candesartan, eprosartan, irbesartan, losartan, olmesartan, telmisartan and valsartan.

Measurements

The measures of interest were individual ARB use over time and switching from the index ARB to an alternative ARB or an ACE inhibitor. Switching from the index ARB to an alternative ARB or an ACE inhibitor was studied during the first ARB treatment episode.

Individual ARB use over time: monthly proportion of total ARB use

The total number of ARB prescriptions was estimated each month during the study period. The proportion of all ARB prescriptions was calculated for each individual ARB.

Individual ARB use over time: quarterly proportion of incident ARB use

Incident ARB use was defined as the first ARB prescription during the study period in a patient with no previous ARB prescription during the preceding 365 days. The total number of patients with incident ARB use was estimated quarterly during the study period. The proportion of all patients with incident ARB use was calculated for each individual ARB.

Switching from the index ARB to an alternative ARB or an ACE inhibitor

A patient was considered to have switched to another ARB if another ARB was prescribed at any time during the first ARB treatment episode. Switching to an ACE inhibitor was only considered in patients that did not switch to another ARB and stopped treatment with the index ARB, to avoid classifying add-on treatment as a switch. The length of the first ARB treatment episode until switching to an alternative ARB, an ACE inhibitor, or treatment discontinuation was estimated for each patient by dividing the total number of tablets by the prescribed daily dose and allowing a maximum gap of 183 days (6 months) between calculated prescription dates. For prescriptions with no recorded daily dose, the median dose for the specific product was used to impute the daily dose or the median dose for all products with the same composition from the same manufacturer, where available, or the defined daily dose [13]. The end date of the treatment episode was the earliest of the end of patient observability (i.e. the last visit of the patient to the practice), database end, 183 days after the calculated end date of the last prescription during the treatment episode, the date of switching to an alternative ARB or the date of switching to an ACE inhibitor. Sensitivity analyses were also carried out, allowing a maximum gap of 90 days or 270 days instead of 183 days.

It was assumed to take up to 6 months after the recall for the switching to occur if patients treated with contaminated ARBs did not switch until the time of the next prescription, as 6 months was the maximum allowed gap between prescription dates during the treatment episode. Hence, a post-recall time period between 11 July 2018 and 20 June 2019 was considered for valsartan (last recall 20 December 2018) [10], 16 July 2018 to 19 August 2019 for irbesartan (last recall 19 February 2019) [10] and 28 December 2018 to 4 September 2019 for losartan (last recall 4 March 2019) [10]. In a sensitivity analysis, the post-recall time period was extended by a further 6 months.

Statistical methods

Prescribing patterns of individual ARBs, as monthly proportion of total ARB use and as quarterly proportion of incident ARB use, were analysed descriptively and by ITS. [14] ITS examines changes in trends before and after an intervention (i.e. the recall) as well as changes in values right before and after the time of the intervention (i.e. the level change). ITS analysis was performed for ARBs that had a minimum proportion of 2% of all ARB use (i.e. candesartan, irbesartan, losartan, olmesartan, telmisartan and valsartan). All analyses were conducted using SAS Enterprise Guide version 7.15.

We also conducted comparative ITS analyses of the difference in ARB proportions [14], comparing valsartan (reference) to candesartan, irbesartan, losartan, olmesartan and telmisartan. July 2018 (the time of the first ARB recall) was selected as an intervention time point for all ITS analyses. A sensitivity analysis was also conducted, using the same number of pre- and post-intervention time points to maximize power (i.e. data from September 2016 for the monthly analysis and data from the first quarter of 2017 for the quarterly analysis) as well as to reduce the influence of historic changes.

Switching was analysed descriptively as a proportion of quarterly ARB treatment episodes leading to a switch during the entire study period. In addition, the rate of switching, defined as the number of ARB treatment episodes leading to a switch per person-year of follow-up, was analysed before and after the ARB recalls, and rates were compared using the z-score test [15,16,17]. The analyses in the study were performed by the authors.

留言 (0)

沒有登入
gif