Quantification of the association between chronic kidney disease status and cause-specific hospitalisation: a population-based cohort study in the UK

Application Number
17_055
Lay Summary

Chronic kidney disease, defined as roughly half of normal kidney function, is common in the general population. People with chronic kidney disease are known to have higher risk of overall hospitalisations (i.e. hospitalisations for any reason) than people without it. However, it remains unclear which causes of hospitalisation are responsible for this increased hospitalisation rate. Previous studies showed that chronic kidney disease is associated with increased risk of a variety of illnesses including heart disease, stroke, infection, bleeding, blood clotting, fracture, and acute kidney disease. Therefore, this study aims to systematically quantify the association between chronic kidney disease status and cause-specific hospitalisation (hospitalisations for common conditions, such as: heart attacks; heart failure; stroke; pneumonia; urinary tract infection; bleeding; deep vein clotting; hip fracture; and acute kidney disease) by comparing cause-specific hospitalisation rates among people with and without chronic kidney disease in absolute and relative terms. Results are important to people with chronic kidney disease and will inform the design of effective preventive care strategies, and future healthcare-service planning.

Technical Summary

We aim to systematically quantify the association between chronic kidney disease (CKD) and cause-specific hospitalisation in absolute and relative terms. We will use a comparative cohort of 242,349 people with CKD (estimated glomerular filtration rate <60 mL/min/1.73m2 twice for >3 months) who were individually matched to a person without known CKD on the same practice register for age, sex and calendar time between 2004 and 2014. Data derive from practices participating in CPRD who have agreed to be linked to Hospital Episode Statistics. Outcomes are hospitalisations for the following common conditions: myocardial infarction; heart failure; cerebral infarction; pneumonia; urinary tract infection; gastrointestinal bleeding; intracranial bleeding; venous thromboembolism; hip fracture; and acute kidney injury. Follow-up continues for each outcome until the first incidence of that outcome after cohort entry, death, renal replacement therapy initiation, change of practice, last data collection, or 31/03/2014. We will estimate an incidence rate difference for each outcome between those with and without CKD, and estimate a relative risk by Cox regression analysis stratifying a matched pair and adjusting for ethnicity, socio-economic and smoking status, BMI, care home residence, and diagnoses of 17 chronic diseases in Quality Outcome Framework. As secondary outcome, we will describe in-hospital mortality after each cause-specific hospitalisation and crudely compare it between matched patients with and without CKD.

Health Outcomes to be Measured

Outcomes are cause-specific hospitalisations for common conditions, based on HES Admitted Patient Care (APC) records.

Collaborators

Dorothea Nitsch - Chief Investigator - London School of Hygiene & Tropical Medicine ( LSHTM )
Masao Iwagami - Corresponding Applicant - University of Tsukuba
Ben Caplin - Collaborator - UCL Hospital
Laurie Tomlinson - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Liam Smeeth - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Sara Thomas - Collaborator - Not from an Organisation

Linkages

HES Admitted Patient Care;Patient Level Index of Multiple Deprivation