Lifestyle and infections : an epidemiological approach on the role of obesity, physical activity, sleep and stress

Abstract

The immune system plays a fundamental role in defending the human body from external pathogens. Some individuals, however, have a weaker immune system leading to more frequent infections. Infections are associated with high burden to society, both in terms of economic costs and morbidity. In this thesis we limited our attention to observational studies to investigate the association between lifestyle and infections. Our findings could be used in support of public health interventions to reduce the susceptibility to infections.

Study I examined the association between physical activity and the occurrence of upper respiratory tract infections (URTI) in a cohort of about 2,000 employed adults. Sleep duration and sleep quality were also investigated in relation to URTI. Participants filled in five questionnaires on demographics and lifestyle and were followed for 9-months for the occurrence of URTI, which was prospectively self-reported through symptoms questionnaires. Due to the excess of individuals with zero infections, hurdle regression models were used to estimate the associations under investigation. During follow-up, 1,597 URTI occurred, but our findings do not support an association of URTI with either physical activity or sleep habits.

Study II investigated the relationship between Body Mass Index (BMI) and infections requiring health care support. Participants in the study are 39,163 Swedish individuals who filled in a questionnaire in autumn 1997 and were followed through record-linkages until December 2016. Infections were identified from the Swedish national inpatient and outpatient registers. Cox proportional hazard models with age as time-scale were fitted to estimate Hazard Ratios and 95% confidence intervals. Extensions of the traditional Cox model, taking into account several infections for each individual, were used in a secondary analysis. Obesity emerged to be a risk factor for infections in both genders. In particular, obese women were at higher risk of skin infections, gastrointestinal tract infections, urinary tract infections and sepsis. Men with obesity showed an increased risk of skin infections. Findings from the extended Cox model were comparable to those from the traditional Cox model.

Study III describes the association between work-related stress and infections. Around 25,000 employed Swedish adults were followed prospectively from September 1997—when they completed a lifestyle questionnaire—until their retirement, emigration, death or December 2016, whichever occurred first. A Swedish version of the Demand-Control Questionnaire was used to assess job stress, whereas infections were identified as in Study II from the National Patient Register. In the main analysis we fitted Cox models accounting for repeated events. We found that higher job demands are associated with increased incidence of infections, in particular upper respiratory tract infections and urinary tract infections. On the other hand, our findings do not support the hypothesis that high job control is associated with a lower occurrence of infections. When combining demand and control dimensions into job strain, we found that workers with active jobs had an increased risk of infections compared to workers with low strain jobs. No difference was observed in workers with high strain jobs compared to those with low strain jobs.

Study IV aimed to explore the relationship between sleep characteristics and inflammatory markers, namely C-Reactive Protein (CRP), Interleukin-6 (IL-6) and Tumor Necrosis Factor (TNF ) in a random sample of 319 non-pregnant women from Uppsala, Sweden. Participants underwent overnight in-home polysomnography (PSG), answered a sleep questionnaire and had blood samples collected the morning after PSG. We first used principal component analysis to reduce the dimensionality of the data and then estimated linear regression (after log-transformation of the outcomes) and quantile regression models to infer on the associations of interest. We found increased CRP levels in women presenting insomnia symptoms (difficulties maintaining sleep or early morning awakenings), whereas sleep duration did not appear to be related with inflammation. From PSG measurements, a reduced REM sleep was associated with higher CRP levels. No association was found with the other markers of inflammation.

Taken together, results from these studies suggest that sleep, stress and obesity might influence the susceptibility to infections, whereas the role of total physical activity is less clear. With these studies we contributed to fill some of the knowledge gaps about the association between lifestyle and infections, but further studies are warranted to overcome the limitations encountered in our research. Furthermore, our studies serve as examples of the wide possibilities offered by statistical tools in the analysis of epidemiological data.