Gout is caused by high levels of urate – a waste product of the body’s metabolism – which deposits as needle-shaped crystals inside joints. These crystals cause gout flares in which there is severe joint pain, swelling and inability to use the joint for 1-2 weeks.
Medicines like allopurinol lower the blood uric acid level and are used to treat gout. However, allopurinol can rarely cause a serious allergic reaction in which a large part of the body’s skin peels off. People with these side effects can get very ill and many of them require hospitalisation. Some people may even die. There is a genetic marker for this allergic reaction, but this is not widely available. In this study, we will develop a method to estimate the risk of this serious side effect in people with gout treated with allopurinol.
We will use anonymized information from Clinical Practice Research Datalink, linked to hospitalisation and mortality records to build and compare different ways of predicting the risk. These prediction models are used to predict whether something of interest will happen to an individual, and if so whether they should have a different management.
Background: Allopurinol, a urate-lowering therapy, is the cornerstone of gout treatment. It can rarely cause severe cutaneous adverse reactions (Stevens-Johnson syndrome/Toxic Epidermal Necrolysis). A prognostic model for this illness has not been developed.
Objectives: To develop and validate a prognostic model for estimating the risk of severe cutaneous adverse reactions due to allopurinol.
Methods: We will use data from the Clinical Practice Research Datalink (CPRD) Aurum and GOLD linked with hospitalisation and mortality records. As these are serious illnesses, they often result in hospital admission and death.
We will use a cohort study design where incident gout patients starting allopurinol for the first time will be followed up from allopurinol initiation until the earliest of severe cutaneous reaction, 100 days follow-up time, death, moving to a different practice, or study end.
Hospitalisation with a discharge diagnosis of generalised skin reaction, Stevens-Johnson syndrome, or toxic epidermal necrolysis or death with any of these conditions recorded as a cause will be the outcome of interest ascertained from hospitalisation and mortality records. A penalized flexible parametric survival model will be used to develop a risk prediction score using CPRD Aurum and with internal validation using bootstrapping to quantify and adjust for optimism. The predictive performance of the score will be validated in CPRD GOLD.
This study will develop a prognostic model to identify people at risk of severe cutaneous adverse reactions from allopurinol. If implemented in clinical practice, it will ensure safer drug prescriptions reducing the risk of these outcomes on patients.
Incident hospitalisation with a discharge diagnosis of severe cutaneous adverse reactions (i.e., Generalized skin eruption due to drugs and medicaments - ICD10 code L27.0; Bullous erythema multiforme/Stevens-Johnson syndrome - ICD10 code L51.1; Toxic epidermal necrolysis/Lyell syndrome - ICD10 code L51.2) or death with any of these conditions as a cause of death within 100 days from allopurinol prescriptions and an algorithm of drug causality for epidermal necrolysis (ALDEN) score≥2 (i.e. there is at least a possible causal association between allopurinol initiation and the severe cutaneous adverse reaction).
Abhishek Abhishek - Chief Investigator - University of Nottingham
Edoardo Cipolletta - Corresponding Applicant - University of Nottingham
Georgina Nakafero - Collaborator - University of Nottingham
Richard Riley - Collaborator - University of Birmingham
HES Admitted Patient Care;ONS Death Registration Data