In this multi-ethnic sample of 1,000 men and women, a greater prevalence of diabetes was found among those with a pathogen burden ≥3, and those with seropositivity to cytomegalovirus, H. pylori, hepatitis A, and herpes simplex virus. These associations were, however, highly confounded; after adjustment for demographic factors, particularly race/ethnicity, the prevalence of diabetes was not related to pathogen burden or to seropositivity to individual pathogens.
We originally hypothesized that the presence of prior infection would influence diabetes prevalence through increasing systemic inflammation [10
]. In this dataset, however, seropositivity to the five pathogens studied was unrelated to several markers of systemic inflammation. Thus, our null results for diabetes are not surprising. As expected, inflammation marker levels were higher among those with diabetes.
Our null findings are contrary to those of several prior studies, which have reported diabetes prevalence to be elevated among participants with antibodies to cytomegalovirus [12
] and herpes simplex virus [13
]. H. pylori
infection has been positively related to diabetes prevalence in most [14
], but not all [18
], prior studies. No literature was identified which assessed the relation between diabetes and infection with C. pneumoniae
, hepatitis A virus, or total pathogen burden. Though not measured in MESA, infection with hepatitis B [19
], hepatitis C [19
], and periodontal pathogens [24
] have generally been adversely associated with type 2 diabetes.
Prior studies exploring associations of infection by the pathogens assessed in MESA to prevalent diabetes are, however, limited. Most had extremely small sample sizes (n
< 150) [12
], the majority had case-control rather than population-based designs [12
], and some were conducted in clinical populations [12
]. Further, as evidenced by our data, the relation between pathogen seropositivity and diabetes may be greatly confounded by sociodemographic factors; it is possible that confounding was not adequately controlled in some of these studies. Finally, publication bias may also provide an explanation for the discrepancy between our results and those of previously published studies.
There are also significant limitations of our study. Foremost, pathogen infection was determined based on seropositivity to IgG antibodies. IgG antibodies reflect prior infection, but are not sensitive indicators of current infection or the chronicity of prior infections. Though our results were null, it is possible that active pathogen infection, or chronic active infection, is associated with systemic inflammation and elevated diabetes risk. Unfortunately, our data are unable to address this issue. Notably, IgG antibodies were used to define infection in most [12
], but not all [14
], prior studies assessing the relation between diabetes and the pathogens studied here.
Regardless of the means by which pathogen infection was assessed, inferences from cross-sectional data exploring the relation between pathogens and diabetes are tenuous, as the temporal direction of the relationship is unclear. While, as proposed in this manuscript, pathogen infection may lead to inflammation and diabetes, an alternate theory suggests that hyperglycemia may impair host defenses and predispose to infection [25
]. Prospective data are clearly needed. This study’s null findings, however, do not provide support for either hypothesis. Another limitation of our study is that the prevalence of seropositivity was high for some pathogens, resulting in relatively low exposure variability.
While acknowledging several limitations, strengths of our study include the relatively large sample size, population-based ascertainment, and the large number of pathogens assessed. In this dataset, pathogen seropositivity was not related to prevalent diabetes or to systemic inflammation. Elevated levels of systemic inflammation markers have, however, emerged as a strong risk factor for diabetes [1
], and rational mechanisms have been identified [10
]. Elucidating factors that contribute to chronic systemic inflammation is an important goal in understanding, and potentially intervening on, type 2 diabetes.