Multiple long-term conditions (MLTC) also known as multimorbidity is defined as the occurrence of several chronic diseases in the same individual. MLTC is considered a global health challenge and has significant negative impact on people’s mental, physical and metabolic health and wellbeing. In addition, MLTC is associated with increased risk of the need for multiple medications, hospital admissions, mortality and increased use of the health care system. Obesity and lifestyle factors are one of the main drivers for MLTC as they are associated with many complications including type 2 diabetes (T2D), heart and brain diseases, certain cancers amongst others.
In this study, we will investigate how different degrees of weight change (-50 to + 50%) affects the disease burden in people living with obesity and MLTC. We will include adults (>18 years) with obesity (BMI ≥ 30 kg/m2) and ≥ 2 obesity related complications as having multimorbidity.
Specifically, we will investigate whether weight change is associated with:
1. Risk of developing additional obesity related complications
2. Risk of developing mental health problems (anxiety, depression, self-harm)
3. Risk of developing substance misuse such as alcohol
4. Mortality
5. Usage of the health care system
6. Number of diseases
7. Number of hospitalisations
8. Number of drugs used
9. Frailty
10. Risk of cardiovascular disease
The findings of this study will be important to patients, health care professionals and payers to evaluate the importance of weight loss and preventing weight gain in reducing the burden of multimorbidity in people living with obesity.
Multiple long-term conditions (MLTC) also known as multimorbidity is associated with increased risk of mortality, polypharmacy and health care resource utilisation (HCRU). Obesity is known as a major driver of MLTC.
This cohort study will investigate the associations between weight change (+/-50%) and risk of incident diseases, mortality, HCRU and polypharmacy. Between January 2001 and December 2017, we will include adults (>18 years) with obesity (BMI ≥ 30 kg/m2), ≥ 2 obesity related complications (ORC) defined as MLTC and ≥1 BMI record during follow-up.
We examine the associations using two designs: 1) weight change over 3 years (change between year 1 and 4) followed by a 4-year follow up and 2) weight change as a continuous measure during 60-months follow-up.
The outcome measures are:
1) Additional ORC’s
2) Mental health problems (Anxiety, depression and self-harm)
3) Substance misuse
4) Mortality
5) HCRU
6) Changes in Disease count
7) Number of hospitalisations
8) Changes in polypharmacy
9) Frailty (eFI: Electronic Frailty Index)
10) QRISK (risk of developing a cardiovascular disease (CVD) over the next 10 years)
We will model time-to-events using Cox proportional hazards from date of inclusion to events or censoring. The effect of weight change will be assessed as a time-varying coefficient.
Potential covariates: comorbidities, medication, BMI at index date, sex, age, alcohol consumption, smoking, ethnicity, socioeconomic status and relevant biomarkers.
Competing risk analysis will ensue for select outcomes. Association between weight change and change in scores will be analysed using mixed models. Statistical analysis will be performed with R (version 4.0.4).
We hypothesize that weight loss will decrease the burden of disease and polypharmacy.
If a weight change is associated with a clinically meaningful change in the selected outcomes, this can guide patients, healthcare professionals, and payers to improve access to evidence-based obesity treatment and prevention.
In time to event analyses (design 1):
1) Risk of developing additional ORCs beyond the exposure/inclusion criteria ORCs. The following comorbidities will be included guided by the study conducted by Kivimaki et al in 2022 (Body-mass index and risk of obesity-related complex multimorbidity: an observational multicohort study, The Lancet Diabetes & Endocrinology. 2022):
• Asthma, back pain, chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), deep vein thrombosis (DVT), dementia, eczema, dyslipidaemia, gastroesophageal reflux disease (GERD), gout, heart failure (HF), hypertension (HT), ischemic heart disease (IHD), musculoskeletal pain including fibromyalgia, non-alcoholic fatty liver disease (NAFLD), non-alcoholic steatohepatitis (NASH), osteoarthritis (hips and knees), obesity related cancers, polycystic ovary syndrome (PCOS), peripheral vascular disease, prediabetes, psoriasis, pulmonary embolism (PE), renal failure, chronic kidney disease (CKD), rheumatoid arthritis (RA) and related disorders, sleep disorders including obstructive sleep apnoea (OSA) and insomnia, stroke and transient ischemic attack (TIA), type 2 diabetes (T2D) and urinary incontinence among females.
2) Risk of developing mental health problems (Anxiety, depression and self-harm)
3) Risk of substance misuse (such as alcohol)
4) Mortality
In analyses with pre-specified follow-up timepoints (design 2):
5) HCRU including number of primary care practitioner (PCP) contacts (phone consultation and face-to-face consultations by nurse of general practitioner), number of clinic contacts, number of prescriptions, length of hospital stay (inpatient and critical care), and number of hospital admissions.
6) Disease count. This is based on the ORC list for outcome 1:
• Asthma, back pain, COPD, CKD, DVT, dementia, eczema, dyslipidaemia, GERD, gout, HF, hypertension, IHD, musculoskeletal pain including fibromyalgia, NAFLD, NASH, osteoarthritis (hips and knees), obesity related cancers, PCOS, peripheral vascular disease, prediabetes, psoriasis, PE, renal failure, CKD, RA and related disorders, sleep disorders including OSA and insomnia, stroke and TIA, T2D and urinary incontinence among females.
Supplemented with
• Bacterial infections and falls
7) Hospitalisations during follow-up (grouped by reason if sample size allows)
8) QRISK (risk of developing a cardiovascular disease (CVD) over the next 10 years)
QRISK is an algorithm which calculates an individual’s 10-year risk of having a heart attack or stroke.
The factors included to identify the patients that have the most risk of a heart disease and stroke are: CKD; migraine; corticosteroids; systemic lupus erythematosus; atypical antipsychotics; severe mental illness; erectile dysfunction; a measure of systolic blood pressure variability.
QRISK is only valid if the patients is not diagnosed already with a coronary heart disease (including angina or heart attack) or stroke/transient ischaemic attack) https://qrisk.org/three/ .
9) Frailty – electronic Frailty Index - a score, identify populations who may be living with varying degrees of frailty).The electronic frailty index (eFI) uses the existing information within the electronic primary health care record to identify populations of people aged 65 and over who may be living with varying degrees of frailty. When applied to a local population it provides opportunity to predict who may be at greatest risk of adverse outcomes in primary care as a result of their underlying vulnerability.
TThe eFI uses existing electronic health records and a ‘cumulative deficit’ model to measure frailty on the basis of the accumulation of a range of deficits. These deficits include clinical signs (e.g. tremor), symptoms (e.g. vision problems), diseases, disabilities and abnormal test values.
It is made up of 36 deficits comprising around 2,000 Read codes. The score is strongly predictive of adverse outcomes and has been validated in around 900,000 patient records.
It presents an output as a score indicating the number of deficits that are present out of a possible total of 36, with the higher scores indicating the increasing possibility of a person living with frailty and hence vulnerability to adverse outcomes.
10) A change in number of drugs and injectables prescribed and change in drug classes
The number of drugs and change in number of drug classes will be applied from a relevant hierarchy, i.e. by unique representations within each BNF chapter. Injectables will be defined by administration route(s).
Camilla Morgen - Chief Investigator - Novo Nordisk A/S
Camilla Morgen - Corresponding Applicant - Novo Nordisk A/S
Abd Tahrani - Collaborator - Novo Nordisk A/S
Christian Kruse - Collaborator - Novo Nordisk A/S
Kamlesh Khunti - Collaborator - University of Leicester
SHWETA UPPAL - Collaborator - Novo Nordisk A/S
Silvia Capucci - Collaborator - Novo Nordisk A/S
Volker Schnecke - Collaborator - Novo Nordisk A/S
HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation