Diabetes is a growing and important health problem. It is well known that diabetes increases the chances of getting some infections and a recent analysis of data collected by GPs in the UK found that the number of infections among older people with diabetes was rising (by 30% between 1997-2010), because diabetes is becoming more common and people in the UK are getting older. Some studies have found that lowering blood sugar levels (through exercise, diet or medication) might prevent some infections or reduce their severity. This is potentially very important in guiding NHS targets for blood sugar levels in people with diabetes, particularly among older people for whom infections may be frequent and potentially serious, and have a large impact of their quality of life. However, lowering blood sugar in older people with diabetes may increase the risk of falls so the ideal blood sugar levels for older people with diabetes are not clear. This project will provide robust estimates of:
1) the risk of a comprehensive range of infections among people with diabetes, compared to those without diabetes, including both older and younger people with diabetes
2) the links between blood sugar control and risks of infection
Diabetes has been linked with increased risk of many infections, but studies have mostly been in small or selected populations, sometimes without gold standard definitions of diabetes, or adequate control for confounding. Most have only considered a small number of infections. There is evidence from population based studies that improved glycaemic control is associated with a significantly lower risk or severity of some infections. This is potentially important in guiding targets for diabetes control, particularly among older people with diabetes for whom infections may be frequent and potentially serious, and for whom there is uncertainty regarding the risk to benefit ratio of improved glycaemic control. We plan to obtain data from Clinical Practice Research Datalink (2008-2015); approximately 85,000 prevalent DM patients aged 40 and over annually. We will link to Hospital Episode Statistics to independently identify serious infections requiring hospitalisation. Infections will be identified by Read codes with simultaneous prescribing of an antibiotic, antifungal or antiviral treatment. In a cohort study design, we will use poisson regression to estimate the rate of specific infections among people with diabetes, compared to those without diabetes, and to relate HbA1c levels from 2008-2010 to future infections (2011-2015). We will also use a self-controlled approach (participants act as their own control; 2008-15).
Health Outcomes to be Measured:
Primary outcomes are episodes of infection amongst the cohort, and secondary outcomes are case-fatality and hospitalisation.
HES Admitted;Patient IMD;Practice IMD (Standard)