In this multiethnic population-based sample, we found that a measure of IB previously associated with vascular disease risk is independently associated with cognitive performance in cross-sectional analyses. We used well-known and validated measures of cognitive function. Cognitive status was assessed at baseline using the 30-point MMSE18
and also during annual follow-up using the TICS-m.19
These results are consistent with prior studies identifying an association between chronic infections and both cognition5–8,21–27
and vascular risk,2–4,28,29
but we further extend prior findings by using a novel weighted average of IB associated with vascular disease risk and by using complementary measures of cognitive function. Our study results point to the possibility that the same cumulative index of pathogens may be a common risk factor for both stroke and cognitive impairment.
Interestingly, we found a more pronounced association of IB with MMSE when treated as a dichotomous variable compared to MMSE as a continuous variable. This might point to a threshold effect or that only a major change in MMSE is associated with a higher IB. The use of the 24-point threshold to define cognitive impairment in Hispanic patients has been a matter of debate since it is not very well documented in this ethnic group,20
but this threshold has been used widely in the literature to describe cognitive impairment and therefore permits comparison with previous studies.
Additionally, we found that the effect of IB on cognitive impairment may depend on certain demographic and vascular risk factors. In general, we found that the magnitude of effects of IB was greater among women, those with lower socioeconomic status (lower levels of education and health insurance), and the physically inactive. The association was most prominently modified by participants' physical activity levels. Among physically inactive participants, a higher IB index was associated with a lower MMSE, while there was no association among physically active participants. This result should be considered exploratory, however, as we investigated multiple interactions. Nevertheless, this observation provides some indirect evidence that the negative effects of chronic infection might be mitigated by beneficial behaviors such as physical activity, and evidence is accumulating that exercise has anti-inflammatory effects.30
IB was not associated, however, with cognitive decline over time, and we did not find any interactions for this with baseline risk factors. The absence of an association with cognitive decline over time could reflect the relatively advanced state of cognitive impairment at the time participants in our relatively elderly cohort were enrolled, limiting our ability to detect further decline with time. Our duration of follow-up, moreover, may have been insufficient to detect a change. Finally, it is possible that a practice effect, such that participants improve scores on cognitive tests over time, may have limited our ability to detect an effect.
Human and animal studies provide evidence that chronic infections with HSV-1, HSV-2, CMV, HIV-1, C pneumoniae
, and H pylori
are associated with an elevated risk of cognitive impairment and different forms of dementia.5–8,21–27
Most of these previous studies were focused on one specific pathogen and the few studies that examined multiple pathogens simultaneously assumed equal weight for each investigated pathogen. For example, one study investigated 400 randomly selected home-dwelling individuals with cardiovascular diseases.5
The primary endpoint was cognitive impairment, defined as MMSE score <24 points at baseline and after 1 year of follow-up. Viral burden at baseline was defined as the number of seropositivities toward HSV-1, HSV-2, and CMV equally divided into 3 categories (0 to 1, 2, or 3). The lowest category was a combination because few people had zero viral seropositivities. Bacterial burden was defined as seropositivities toward C pneumoniae
and Mycoplasma pneumoniae
divided into 3 categories (0, 1, or 2). Logistic regression was performed with the categories of viral or bacterial burden, various risk factors, and demographic data as independent variables. In subjects with 3 viral seropositivities (compared to 1), the adjusted hazard ratio for cognitive impairment at baseline was 2.5 (1.3–4.7) and after 1 year it was 2.3 (1.2–4.6). However, no association was observed between cognition and bacterial burden. Another investigator studied a subset (n = 1,204) of the participants in the Sacramento Area Latino Study on Aging.24
Baseline serum samples were assayed for levels of immunoglobulin G antibodies to CMV and HSV-1. Participants were screened annually over a 4-year period for cognitive function (by MMSE) and episodic memory (by Delayed Recall Scale scores). They found a higher rate of cognitive decline over the 4-year period in subjects with the highest CMV antibody levels at baseline than in individuals with the lowest levels (p
= 0.003). Unlike these studies, we did not assume that each infection should be associated with a similar risk of cerebrovascular disease or cognition, and we implemented a novel weighted average of IB that was previously associated with cerebrovascular events. However, since previous studies showed associations primarily with viruses, we also performed an exploratory analysis to assess whether the driving force behind the association of IB index with cognition might be the viral pathogen burden. The estimates for the associations with cognition using viral serologies alone were almost identical to those using a combined bacterial and viral score, providing support for the notion that most of the effect of IB on stroke and cognition is mediated through viral rather than bacterial serologies.
The mechanism for this association remains uncertain. It may be that chronic infection due to these pathogens contributes to the overall inflammatory milieu and, together with other risk factors, leads to atherosclerosis, subclinical stroke, and dementia. Inflammation in brain vessels has been postulated to play an important role in both vascular dementia and AD.31
In addition, a direct toxic effect of some neurotropic agents may play a role in the development of cognitive impairment.32,33
Further studies are required to establish the pathogenic mechanisms.
In our univariate analysis, we found an association of IB index with insurance status and education, both proxies for socioeconomic status (SES), as well as race-ethnicity. In the multivariate analysis after adjusting for these demographic variables, the association of IB index with cognition was attenuated. It might be that the known socioeconomic gradients concerning the incidence of cognitive impairment34
are partly explained by a higher IB.35
For example, data from the National Health and Nutrition Examination Survey III addressed socioeconomic and race-ethnic differences in infection status in the United States.36
, CMV, HSV-1, and hepatitis B virus were used for the infection burden analyses.36
The authors found that individuals with less than high school education had roughly 50% increased risk of having an additional infection compared with high school graduates, whereas those with postsecondary education had 50% lower odds.36
Income had similar effects, whereby low income was associated with 33% higher odds of an additional infection, and high income with 45% lower odds, compared with the middle-income group.36
This association might be due to a higher exposure rate but it might also be due to an initial susceptibility to infections triggered by socioeconomic stressors. There is evidence of an association between low SES, increased stress, and enhanced susceptibility to several viruses.37
Our study has limitations. Because it is cross-sectional, we cannot make conclusions about the direction of associations. The infections we investigated, however, most likely preceded the development of cognitive impairment, since the antibody pattern reflects chronic infectious status and representative studies have shown that the majority of CMV and HSV-1 infections occur in childhood.38,39
Second, it was not possible to examine the relationship between infection and specific forms of cognitive impairment. For example, earlier research has shown that CMV DNA is found in a higher proportion of brains of people with vascular dementia than in age-matched controls.7
Third, we did not use detailed neuropsychological testing in our analyses. However, both the MMSE as well as the TICS-m are well-established tools to assess cognitive performance in large cohorts. Fourth, we did not have APOE
genotype information in all subjects. However, in the subgroup analyses, we found that IB index was independently associated with cognition even after adjusting for APOE
genotype, and the effects of IB index on cognition did not differ by APOE
genotype. These findings should be interpreted with caution due to potential selection bias. Finally, a few factors may have reduced the ability to detect a strong association between IB index and cognitive decline over time in our study. For example, a moderate practice effect over the follow-up period might have played a role, or a threshold effect, indicating that when the damage is already done there is no further decline.
The strengths of this study include the population-based multiethnic cohort, including a large proportion of Hispanic subjects, who are frequently underrepresented in studies of cognition, and the ability to adjust for numerous potential covariates. In addition, GEE analyses were used to model corresponding trajectories of cognitive decline.
Our results need to be validated in independent populations before they can be generalized. If confirmed, however, these findings could have potential clinical implications. For example, treatment and eradication of these pathogens might have a positive impact on cognition as well as stroke, thereby addressing 2 major causes of neurologic disease burden worldwide. In the case of viral pathogens, early childhood vaccination or antiviral treatment could decrease stroke risk and cognitive impairment. Based on our data, early intervention might be more promising, since there was no association with decline over time in this elderly cohort.
Our results extend the findings of previous studies aimed at investigating the association of chronic infections and cognitive performance and specifically point to IB as a common risk factor of both stroke and cognitive impairment.