Diabetes and infection: assessing the association with glycaemic control in population-based studies

Summary

Diabetes is a leading cause of morbidity and mortality. The global burden of diabetes is rising because of increased obesity and population ageing. Although preventive and treatment measures are well documented for macrovascular and microvascular complications, little such guidance exists for infections in people with diabetes, despite evidence suggesting greater susceptibility to infections, and worse outcomes. In particular, few studies have characterised the relation between glycaemic control and infectious disease, which we discuss in this Review. Some large population-based observational studies have reported strong associations between higher HbA_1c and infection risks for both type 1 and type 2 diabetes. However, studies are contradictory, underpowered, or do not control for confounders. Evidence suggests that better glycaemic control might reduce infection risk, but further longitudinal studies with more frequent measures of HbA_1c are needed. Older people (aged 70 years or older) with diabetes are at increased risk of complications, including infectious diseases. There is more uncertainty about appropriate glycaemic control targets in this age group, and evidence suggests that glycaemic control is often neglected. Robust evidence from cohorts with sufficient numbers of older people would help to develop clinically relevant guidelines and targets to reduce mortality, morbidity, and antibiotic use, and to improve quality of life.

Introduction

Diabetes is one of the leading causes of morbidity worldwide, and creates a huge burden of disease. Diabetes is associated with a range of severe acute and chronic complications, which results in a correspondingly large financial burden. It has been estimated that US$612 billion are spent annually on diabetes care, accounting for 11% of total adult health spending.¹ In high-income countries, diabetes prevalence increases steeply with age,² and more than 25% of older adults (throughout this Review older people refers to those aged 70 years or older) in British care homes might have diabetes.³

A relation between diabetes and infection has long been accepted clinically. Some rare infections, known as signal infections4, 5 (severe infections which are not self-limiting and are likely to be diagnosed irrespective of differences in physician behaviour or previous diagnosis of diabetes), are often pathognomonic of a patient having diabetes (panel 14, 5, 6).7, 8 Examples of signal infections include emphysematous pyelonephritis, malignant otitis externa,⁹ mucormycosis,¹⁰ and Fournier's gangrene.⁸ Although other infections, such as lower respiratory tract infections, urinary tract infections, and cellulitis, are very common in the general population, people with diabetes are at higher risk.⁶ A key primary care study in the Netherlands⁶ prospectively followed up 7000 patients with diabetes for 1 year, comparing their attendances for infection with those of control patients with hypertension. Results of this investigation reported that patients with type 1 diabetes and type 2 diabetes were at high risk of lower (but not upper) respiratory tract infections (type 1 diabetes: adjusted odds ratio [aOR] 1·42 [95% CI 0·96–2·08], type 2 diabetes: 1·32 [95% CI 1·13–1·53]), and urinary tract infections (type 1 diabetes: aOR 1·96 [95% CI 1·49–2·58], type 2 diabetes: 1·24 [95% CI 1·10–1·39]). A study¹¹ in the USA estimated that 10% of Emergency Department visits in patients with diabetes resulted from infection, and patients with diabetes were twice as likely to be admitted to hospital with infection as were patients without diabetes. Results of a population-based study¹² from Canada reported that almost half of the patients with diabetes had an infection that led to hospital admission or a physician insurance claim over a 1 year period.

Infections can be particularly common and more serious in older people with type 2 diabetes.¹³ The rise in diabetes prevalence is most alarming in older people in high-income countries, and an analysis¹⁴ of primary care data in the UK in older patients with diabetes noted that the burden of diabetes-associated infections rose by 30% between 1997 and 2010, with lower respiratory tract infection having the highest incidence (152·7/1000 person-years). This substantial increase was attributed to both the rising prevalence of type 2 diabetes, and population ageing.¹⁵ Furthermore, evidence suggests that diabetes care and targets are often neglected in older people.2, 16 The 2015 draft NICE guidance for type 2 diabetes¹⁷ did not include any management or prevention strategy for infection, nor did a position statement on diabetes in older people.¹⁸

Evidence from a meta-epidemiological review¹⁹ of 97 prospective cohort studies (820 900 people with diabetes) that assessed the risk of cause-specific mortality reported a substantial increased risk of infection-related mortality when excluding pneumonia (relative risk [RR] 2·39, 95% CI 1·95–2·93; 1083 deaths) and of pneumonia mortality itself (RR 1·65, 95% CI 1·45–1·92).¹⁹ Results showed an increased risk for death from all causes in patients with diabetes, compared with people without diabetes (RR 1·80, 95% CI 1·71–1·90). This high all-cause mortality risk is not supported by the higher quality evidence from randomised controlled trials, but this might be because these epidemiological studies included older patients, and more patients with comorbidities or diabetes complications, who are frequently excluded from the randomised controlled trials. Panel 2 outlines in more detail the relation between diabetes and infections in older people.

Randomised controlled trials have suggested microvascular and macrovascular benefits from improved glycaemic control in younger, fitter people.26, 27, 28 Randomised controlled trials29, 30, 31 that included older patients (over 60 years), with a longer history of diabetes and at higher cardiovascular risk, did not find convincing evidence of any benefits on all-cause cardiovascular disease mortality.²⁰ Reviews and editorials on this topic have concluded that “further assessment of the risks and benefits (of tight glucose control) [in all people with diabetes] is therefore needed”.20, 32 Trial evidence thus far has led to well documented and accepted targets for glycaemic control, and national guidelines for prevention of complications in people with diabetes.¹⁷ For example, the NICE guidelines¹⁷ for glycaemic control focus on macrovascular and microvascular outcomes, and do not mention the potential association between glycaemic control and infection.

Infectious disease risks have been neglected, and have not been reported by most of these trials, despite immunological evidence (panel 3) that hyperglycaemic states negatively affect outcomes during infections in patients with diabetes.33, 34, 35 Patients with diabetes are not only more susceptible to infection but often have different disease courses once the infection has been established, which is reflected by increased rates of hospital admission,⁴³ length of stay,44, 45 and complications.⁴⁶ Therefore, infections in older patients with diabetes could have at least as much effect on their quality of life as macrovascular and microvascular risks.

Other review articles7, 51, 52 have considered the association between diabetes and infection, but none have presented a comprehensive review of the effects of glycaemic control on infectious disease outcomes. In this Review, we examine the mainly observational evidence from population-based epidemiological studies that assess the association between infection risk and glycaemic control in type 1 and type 2 diabetes, and the potential effect of ageing on this association.