In recent years, there has been a significant increase in the use of opioids (potent painkillers) for chronic pain. However, this has raised concerns of potential harmful effects associated with its long-term use, such as dependence, addiction, bone fractures, and death. In response to this, national healthcare organisations in the United Kingdom (UK) have released guidelines, advocating healthcare providers to consider reducing or stopping long-term opioid therapy when benefits are minimal and outweighed by potential harms.
To help healthcare providers make informed decisions, we need to find out how patients reduced or stopped their long-term opioid use, and what factors influence these actions. We also want to explore any potential risks that may arise following these recommendations. Such information will help healthcare providers to figure out target patients, for whom discontinuation/reduction of long-term opioid therapy is likely and beneficial.
Using UK primary care data, hospitalisation record, and national death information from the Clinical Practice Research Datalink, we will focus on how opioids are used in patients with chronic pain. In particular, whether patients discontinued or reduced their long-term opioid therapy. We will then investigate various factors that trigger these prescribing changes. Additionally, we will use statistical methods to compare the occurrence of negative consequences in this group of patients with those who continued their long-term opioid treatment.
The findings of this research will provide evidence to guide safer prescribing practices for opioids, inform the development of strategies to minimise potential harms of long-term opioid therapy, and ultimately enhance patient safety.
The increased use of opioids for chronic non-cancer pain (CNCP) in the UK raises serious risks, such as addiction, respiratory depression, bone fractures, and drug-related death. Recent UK NICE guidelines advocate for reducing or discontinuing opioid use when significant harms are present. To guide practice, we aim to assess factors influencing the reduction or discontinuation of long-term opioid therapy (LTOT) and compare risks of serious adverse events associated with these actions among patients with CNCP in England.
This study comprises three work packages (WP). WP1 is a drug utilisation study that will use data from CPRD Aurum from 1 March 2000 to 31 March 2021. The cohort includes CNCP patients who received ≥1 prescription for opioids in their entire record. Patients will be followed from the initiation of opioid therapy until they transfer out of their GP practice, experience death, or reach the end of study period, whichever occurs first, to identify opioid prescribing patterns and the prevalence of LTOT. WP2 is a nested case-control study that will use CNCP patients classified as long-term opioid users from WP1 and group them as discontinued/reduced LTOT vs. continued LTOT. Predictors of these two treatment changes will be analysed using mixed-effects logistic regression models. WP3 will use data from CPRD linked with HES and ONS registry to assess the incidence of bone fractures and opioid-related death in patients discontinued/reduced LTOT vs. those continued LTOT. We will use a 1:1 (or 1: N) propensity score matching method to match the samples. Then, we will use marginal structural Cox regression with inverse probability of treatment weighting, adjusting for covariates, to determine associations between LTOT discontinuation/reduction and the time to events. The proportional hazards assumption will be evaluated by plotting Schoenfeld residuals. We will also use parametric survival models as alternative analytical approaches.
The following outcomes will be measured for each individual work package:
1. Work package 1- the primary outcomes are patterns of opioid utilisation and the prevalence of long-term opioid therapy (LTOT). Patterns of opioid utilisation include:
• Opioid characteristics, stratified by potency (strong vs weak), formulation (immediate release vs modified release), and duration (long-term vs no).
• Mean or median daily dosage (will be converted into oral morphine equivalent dose [OMEQ] based on the conversion factors (1), and trichotomised as low [<50 mg OMEQ] vs medium [50-120 mg OMEQ] vs high [>120 mg OMEQ])
• Median annual days covered (will be calculated from the sum of days that each person was covered by any opioid within a given patient-year)
• Concomitant medication utilisation (benzodiazepines, gabapentinoids, antidepressants, Z-drugs, antipsychotics and muscle relaxants, defined as a filled prescription at any time during the long-term opioid use or in the 7 days before it begins)
• Time to the initiation of opioids post CNCP diagnosis.
Currently this is no consensus on the definition of LTOT. Therefore, we propose a definition for LTOT based on existing literature (25). The suggested definition is receiving three or more opioid prescriptions over a period of 90 days or longer in any consecutive 180-day period, with no 28-day gap between fill dates. However, it is important to acknowledge that this definition remains subject to potential revision based on the analysis of opioid utilisation patterns derived from the datasets. Therefore, we will explore potential alternative definitions that may become available, or we may devise our own after examining the data.
2. Work package 2 - we will examine two co-primary outcomes:
• The discontinuation of LTOT.
• The reduction of LTOT.
(Specific definitions of these two outcomes are provided in section L. Briefly, the discontinuation of LTOT is defined as the absence of any opioid prescription claims for at 180 days following any period after a continuous LTOT period of 90 days or more. The reduction of LTOT is defined as a ≥ 10% reduction in average OMEQ compared with the baseline dose, in both of two consecutive 90-day periods during the follow-up period.)
Secondary outcomes of interest include:
• Time to the first treatment changes (i.e., LTOT discontinuation or reduction) after the initiation of opioids.
• The proportion of patients re-initiated opioid use after the first treatment break (i.e., LTOT discontinuation) and the time from it to the re-initiation of opioids.
3. Work package 3 - we will examine two co-primary outcomes:
• Bone fractures following the first discontinuation/reduction of LTOT.
• Opioid related death following the first discontinuation/reduction of LTOT.
Evangelos Kontopantelis - Chief Investigator - University of Manchester
Qian Cai - Corresponding Applicant - University of Manchester
Charlotte Morris - Collaborator - University of Manchester
Christos Grigoroglou - Collaborator - University of Manchester
Darren Ashcroft - Collaborator - University of Manchester
Douglas Steinke - Collaborator - University of Manchester
Li-Chia Chen - Collaborator - University of Manchester
Louise Gorman - Collaborator - University of Manchester
Thomas Allen - Collaborator - University of Manchester
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Rural-Urban Classification