381 research outputs found
Investigating effectiveness and safety of sodium-glucose co-transporter 2 inhibitors in type 2 diabetes in Scotland
No full textSodium-glucose co-transporter 2 inhibitors (SGLT2i) are licensed for the treatment of type 2 diabetes (T2D).
This pharmacoepidemiology PhD project investigates the effect of SGLT2i exposure in people
living with T2D in Scotland, using the national patient record (Scottish Care Information
(SCI)-Diabetes) which is linked to other datasets, using an observational study design.
This database is ready-provisioned for research use as the Scottish Diabetes Research
Network Epidemiology (SDRN-Epi) database.
In this project, I explore the effect of SGLT2i exposure on continuous variable outcomes,
on binary clinical event outcomes for which a beneficial effect might be expected and on
potential safety outcomes.
Despite SGLT2i efficacy in large cardiovascular outcome trials, there remains uncertainty
about whether the treatment effect described in the randomised control trials (RCTs)
extends to the population with a broader baseline cardiovascular disease (CVD) risk
encountered in day-to-day practice compared to those included in the trials. There are also
questions about whether the treatment effects persist in the real-world and over a longer
follow-up to that studied in the RCTs. Observational pharmacoepidemiology can explore
the effect of exposure in these areas.
For assessing drug safety, observational pharmacoepidemiology studies remain the mainstay
for investigation, as safety signals from RCTs are frequently uncertain (and they are
often not powered to detect harm).
In Chapter 1 (Introduction), I present a paper entitled “Pharmacoepidemiology: Using
randomised control trials and observational studies in clinical decision-making. I compare
and contrast RCTs and observational pharmacoepidemiology studies and, I describe why, despite possibly being biased and confounded, observational pharmacoepidemiology
studies still have merit.
In Chapter 2 (Background), I show a systematic review article entitled: “Sodium–Glucose
Co-Transporter 2 Inhibitors (SGLT2i) Exposure and Outcomes in Type 2 Diabetes: A
Systematic Review of Population-Based Observational Studies”. I undertook this review
to explore studies similar to those I planned to undertake, and to understand the areas of
uncertainty that needed to be explored.
In Chapter 3 (Methods), I describe the cohort study design with prospective follow-up and
the methodological approach employed in detail and the statistical analysis plan.
For continuous variable outcomes, I describe an exposed-only, self-controlled study using
linear mixed effect models. In these models, people act as their own controls (which cancels
out the effect of time-invariant confounding) and models are fitted with an autoregressive
correlation structure to reduce the effect of regression-to-the-mean and autoregressive
drift.
For binary clinical event outcomes, I explain how I employed different Poisson models in a
cohort study with prospective follow-up to evaluate the effect of drug exposure.
I then explain the difficulties that arose as I developed the pharmacoepidemiology pipeline
and the methodological changes that deviated from the original statistical analysis plan I
had developed and why.
In Chapter 4 (Results: The association of SGLT2i exposure on continuous variable
outcomes - a within-person, exposed-only cohort study), I illustrate the effect of SGLT2i
exposure on a variety of biomarker outcomes. For glycosylated haemoglobin (HbA1c), I
show that SGLT2i exposure was associated with a profound and sustained reduction in this
variable (max -7.90 mmol/mol, 95% confidence interval (CI) -10.50, -5.28 at >102 months
post-exposure). For systolic blood pressure, I show that exposure was associated with
v
a sustained fall, up to 90-months post-exposure (max -2.28 mmHg, 95% CI -2.47, -2.09
at 18-24 months). For body mass index, exposure was associated with a sustained fall
compared to pre-exposure time (max -0.52 kg/m2, 95% CI -0.54, -0.49 at 30-36 months).
For estimated glomerular filtration (eGFR), SGLT2i exposure was associated with an early
fall but then a protective association further out in time (max fall, -0.54 ml/min/1.73 m2, 95%
CI -0.59, -0.48). For blood lipids, SGLT2i exposure was associated with a very small and
likely clinically insignificant changes.
In Chapter 5 (Results: The association of SGLT2i exposure on binary clinical event
outcomes - a cohort study), I explore the effect of SGLT2i exposure using Poisson
models.
Outcomes for which a beneficial effect might be expected: for hospitalisation for heart
failure, SGLT2i exposure was associated with a reduction in admissions in both the
adjusted exposed/unexposed model (risk ratio (RR) 0.73, 95% CI 0.65, 0.82) and a
reduction in risk that faded-in at half-maximal effect ~8 weeks post-exposure, in the
fade-in Poisson model (RR 0.54, 95% CI 0.38, 0.75). For the remaining outcomes for
which a beneficial effect might be expected, the exposure/outcome relationship best
supported by the data was the adjusted exposed/unexposed model (which includes a
possible instantaneous effect of exposure on the outcome). For PVD, although the primary
effect estimate of interest did not return a positive association, the lower bound of the
95% CI were approaching one in both the adjusted exposed/unexposed and the inverseprobability
of treatment weighting (IPTW) model; in the subgroup with preserved renal
function, the model suggests an associated increased risk of this outcome.
Potential adverse events: For diabetic ketoacidosis (DKA), the best fitting duration-response
Poisson model appeared to be for a cumulative effect of exposure being associated with
a lower risk of this outcome. On closer inspection of the data, it was clear that there
was an apparent transient increased risk of this outcome followed by an lowering of risk,
a depletion of the vulnerable effect, where those destined developed DKA did so early
after exposure and had SGLT2i therapy stopped. For lower limb amputation (LLA), the
primary efficacy parameter of interest suggested no association with SGLT2i exposure
and LLA. It should be noted however that the lower bound of the 95% CI for in both the
adjusted exposed/exposed and IPTW was approaching one. In those with reduced renal
function, there was an association of increased risk in the adjusted exposed/unexposed
model for LLA. For necrotising fasciitis in the whole population there was no association
of increased risk but in those with reduced renal function, the IPTW model suggested
increased risk. SGLT2i exposure did not appear to be associated with the following
safety outcomes: hospitalisations for fractures, bladder cancer, breast cancer, urinary tract
infections or hypoglycaemia.
In Chapter 6 (Discussion), I summarise the findings of my studies and the important
real-world confirmation and reassurance of efficacy on major cardiovascular outcomes they
provide. I describe the current regulatory positions regarding SGLT2i and the literature
about these medicines. I discuss in detail the difficulties of modelling exposure-outcome
effects whilst minimising important potential biases and confounding, the challenge of
establishing a generic pharmacoepidemiology pipeline and the lessons learned. I set
out the next steps for this research, including the framework of a planned international
meta-analysis following the methods already described and the other exposures to be
studied (including the individual component members of the SGLT2i class)
Dataset pertaining to the publication "Quantitative levels of serum N-glycans in type 1 diabetes and their association with kidney disease"
No full textThe files are comma separated and contain measurements for 46 N-glycan peaks on 1581 serum samples (corresponding to 1565 unique participants) from SDRNT1BIO and from a pool of healthy controls. N-glycan peaks are expressed as absolute levels in ng/µL in volume of serum and as percentage areas (in this case, result files also contain a Man3 column). File 'sample_script.R' contains code to read these files, recompute percentage areas after exclusion of Man3, and derivation of 18 summary measures that aggregate the sums of peaks having features related to antennary structure, galactosylation, fucosylation, or sialylation.Shehni, Akram Asadi; Wilkinson, Hayden; Blackbourn, Luke AK; Colombo, Marco; Saldova, Radka; Colhoun, Helen M. (2020). Dataset pertaining to the publication "Quantitative levels of serum N-glycans in type 1 diabetes and their association with kidney disease", [dataset]. University of Edinburgh. https://doi.org/10.7488/ds/2856
Health Survey for England 1994:a survey carried out on behalf of the Department of Health
No full textHealth Survey for England 1994:a survey carried out on behalf of the Department of Health
No full textEfficacy and safety of adding alirocumab to rosuvastatin versus adding ezetimibe or doubling the rosuvastatin dose in high cardiovascular-risk patients:The ODYSSEY OPTIONS II randomized trial
No full textOBJECTIVE: To compare lipid-lowering efficacy of adding alirocumab to rosuvastatin versus other treatment strategies (NCT01730053).METHODS: Patients receiving baseline rosuvastatin regimens (10 or 20 mg) were randomized to: add-on alirocumab 75 mg every-2-weeks (Q2W) (1-mL subcutaneous injection via pre-filled pen); add-on ezetimibe 10 mg/day; or double-dose rosuvastatin. Patients had cardiovascular disease (CVD) and low-density lipoprotein cholesterol (LDL-C) ≥70 mg/dL (1.8 mmol/L) or CVD risk factors and LDL-C ≥100 mg/dL (2.6 mmol/L). In the alirocumab group, dose was blindly increased at Week 12 to 150 mg Q2W (also 1-mL volume) in patients not achieving their LDL-C target. Primary endpoint was percent change in calculated LDL-C from baseline to 24 weeks (intent-to-treat).RESULTS: 305 patients were randomized. In the baseline rosuvastatin 10 mg group, significantly greater LDL-C reductions were observed with add-on alirocumab (-50.6%) versus ezetimibe (-14.4%; p < 0.0001) and double-dose rosuvastatin (-16.3%; p < 0.0001). In the baseline rosuvastatin 20 mg group, LDL-C reduction with add-on alirocumab was -36.3% compared with -11.0% with ezetimibe and -15.9% with double-dose rosuvastatin (p = 0.0136 and 0.0453, respectively; pre-specified threshold for significance p < 0.0125). Overall, ∼80% alirocumab patients were maintained on 75 mg Q2W. Of alirocumab-treated patients, 84.9% and 66.7% in the baseline rosuvastatin 10 and 20 mg groups, respectively, achieved risk-based LDL-C targets. Treatment-emergent adverse events occurred in 56.3% of alirocumab patients versus 53.5% ezetimibe and 67.3% double-dose rosuvastatin (pooled data).CONCLUSIONS: The addition of alirocumab to rosuvastatin provided incremental LDL-C lowering versus adding ezetimibe or doubling the rosuvastatin dose.</p
A scalable formulation of joint modelling for longitudinal and time to event data and its application on large electronic health record data of diabetes complications
No full textINTRODUCTION:
Clinical decision-making in the management of diabetes and other chronic diseases depends upon individualised risk predictions of progression of the disease or complica- tions of disease. With sequential measurements of biomarkers, it should be possible to make dynamic predictions that are updated as new data arrive. Since the 1990s, methods have been developed to jointly model longitudinal measurements of biomarkers and time-to-event data, aiming to facilitate predictions in various fields.
These methods offer a comprehensive approach to analyse both the longitudinal changes in biomarkers, and the occurrence of events, allowing for a more integrated understanding of the underlying processes and improved predictive capabilities. The aim of this thesis is to investigate whether established methods for joint modelling are able to scale to large-scale electronic health record datasets with multiple biomarkers measured asynchronously, and evaluates the performance of a novel approach that overcomes the limitations of existing methods.
METHODS:
The epidemiological study design utilised in this research is a retrospective observa- tional study. The data used for these analyses were obtained from a registry encompassing all individuals with type 1 diabetes in Scotland, which is delivered by the Scottish Care Information - Diabetes Collaboration platform. The two outcomes studied were time to cardiovascular disease (CVD) and time to end-stage renal disease (ESRD) from T1D diag- nosis. The longitudinal biomarkers examined in the study were glycosylated haemoglobin (HbA1c) and estimated glomerular filtration rate (eGFR). These biomarkers and endpoints were selected based on their prevalence in the T1D population and the established association between these biomarkers and the outcomes.
As a state-of-the-art method for joint modelling, Brilleman’s stan_jm() function was evaluated. This is an implementation of a shared parameter joint model for longitudinal and time-to- event data in Stan contributed to the rstanarm package. This was compared with a novel approach based on sequential Bayesian updating of a continuous-time state-space model for the biomarkers, with predictions generated by a Kalman filter algorithm using the ctsem package fed into a Poisson time-splitting regression model for the events. In contrast to the standard joint modelling approach that can only fit a linear mixed model to the biomarkers, the ctsem package is able to fit a broader family of models that include terms for autoregressive drift and diffusion. As a baseline for comparison, a last-observation-carried-forward model was evaluated to predict time-to-event.
RESULTS:
The analyses were conducted using renal replacement therapy outcome data regarding 29764 individuals and cardiovascular disease outcome data on 29479 individuals in Scotland (as per the 2019 national registry extract). The CVD dataset was reduced to 24779 individuals with both HbA1c and eGFR data measured on the same date; a limitation of the modelling function itself. The datasets include 799 events of renal replacement therapy (RRT) or death due to renal failure (6.71 years average follow-up) and 2274 CVD events (7.54 years average follow-up) respectively. The standard approach to joint modelling using quadrature to integrate over the trajectories of the latent biomarker states, implemented in rstanarm, was found to be too slow to use even with moderate-sized datasets, e.g. 17.5 hours for a subset
of 2633 subjects, 35.9 hours for 5265 subjects, and more than 68 hours for 10532 subjects. The sequential Bayesian updating approach was much faster, as it was able to analyse a dataset of 29121 individuals over 225598.3 person-years in 19 hours. Comparison of the fit of different longitudinal biomarker submodels showed that the fit of models that also included a drift and diffusion term was much better (AIC 51139 deviance units lower) than models that included only a linear mixed model slope term. Despite this, the improvement in predictive performance was slight for CVD (C-statistic 0.680 to 0.696 for 2112 individuals) and only moderate for end-stage renal disease (C-statistic 0.88 to 0.91 for 2000 individuals) by adding terms for diffusion and drift. The predictive performance of joint modelling in these datasets was only slightly better than using last-observation-carried-forward in the Poisson regression model (C-statistic 0.819 over 8625 person-years).
CONCLUSIONS:
I have demonstrated that unlike the standard approach to joint modelling, implemented in rstanarm, the time-splitting joint modelling approach based on sequential Bayesian updating can scale to a large dataset and allows biomarker trajectories to be modelled with a wider family of models that have better fit than simple linear mixed models. However, in this application, where the only biomarkers were HbA1c and eGFR, and the outcomes were time-to-CVD and end-stage renal disease, the increment in the predictive performance of joint modelling compared with last-observation-carried forward was slight. For other outcomes, where the ability to predict time-to-event depends upon modelling latent biomarker trajectories rather than just using the last-observation-carried-forward, the advantages of joint modelling may be greater.
This thesis proceeds as follows. The first two chapters serve as an introduction to the joint modelling of longitudinal and time-to-event data and its relation to other methods for clinical risk prediction. Briefly, this part explores the rationale for utilising such an approach to manage chronic diseases, such as T1D, better. The methodological chapters of this thesis describe the mathematical formulation of a multivariate shared-parameter joint model and introduce its application and performance on a subset of individuals with T1D and data pertaining to CVD and ESRD outcomes.
Additionally, the mathematical formulation of an alternative time-splitting approach is demonstrated and compared to a conventional method for estimating longitudinal trajectories of clinical biomarkers used in risk prediction. Also, the key features of the pipeline required to implement this approach are outlined. The final chapters of the thesis present an applied example that demonstrates the estimation and evaluation of the alternative modelling approach and explores the types of inferences that can be obtained for a subset of individuals with T1D that might progress to ESRD. Finally, this thesis highlights the strengths and weaknesses of applying and scaling up more complex modelling approaches to facilitate dynamic risk prediction for precision medicine
Efficacy and safety of alirocumab in individuals with type 2 diabetes mellitus with or without mixed dyslipidaemia: Analysis of the ODYSSEY LONG TERM trial
No full textAlirocumab, a monoclonal antibody to proprotein convertase subtilisin/kexin type 9, significantly reduces low-density lipoprotein cholesterol (LDL-C). We evaluated the efficacy and safety of alirocumab in individuals with type 2 diabetes mellitus (T2DM) with versus without mixed dyslipidaemia (MDL, defined as baseline LDL-C ≥70 mg/dL [1.8 mmol/L] and triglycerides ≥150 mg/dL [1.7 mmol/L])
Considerations for more actionable consensus guidance for monitoring individuals with islet autoantibody‑positive pre‑stage 3 type 1 diabetes. Reply to Mallone R [letter]
No full text- …
