Infections May Play a Role in Cognitive Decline, Dementia

Abstract & Commentary

By Richard R. Watkins, MD, MS, FACP, Division of Infectious Diseases, Akron General Medical Center, Akron, OH; Associate Professor of Internal Medicine, Northeast Ohio Medical University, Rootstown, OH. Dr. Watkins reports no financial relationships relevant to this field of study.

This article originally appeared in the May 2013 issue of Infectious Disease Alert.

Synopsis: This study reports an association between infection burden and dementia, although the relationship may not be causal.

Source: Katan M, et al. Infectious burden and cognitive function: The Northern Manhattan Study. Neurology 2013;80: 1209-1215.

Do latent infections cause dementia? While certain chronic illnesses (e.g., peptic ulcer disease, Burkitt’s lymphoma, and cervical cancer) have known infectious etiologies, mainstream research has not elucidated a significant role for microbes in cognitive decline. But emerging data suggest an association between some viruses and bacteria and Alzheimer dementia (AD). Moreover, the detrimental neurocognitive effects of HIV infection are well established. Dementia represents an enormous financial burden on the health care system, comparable to heart disease and cancer.1 Katan and colleagues present epidemiologic evidence of an association between infectious burden (IB) and dementia from a large multiethnic cohort.

The Northern Manhattan Study (NOMAS) was a multiethnic stroke-free cohort that enrolled 3298 participants ≥ 40 years of age between 1993 and 2001. Of these, 1625 subjects (65% women, mean age 69 years, 58% Hispanic) had serologic measurements taken for Chlamydia pneumoniae, Helicobacter pylori, cytomegalovirus, and herpes simplex virus type 1 and type 2. Cognitive status was assessed at baseline using the mini-mental state exam (MMSE) and then annually by a telephone interview. A subset also had the number of APOE ε4 alleles (known to increase risk for AD) identified. IB was used as the main predictor, which was determined by the relationship of individual serologic test results to the risk of stroke using estimates from Cox proportional hazard models. In a post hoc analysis, the investigators created a viral burden index (VIB) and also tested its association with cognition in the same manner as the overall IB index.

The results of the study were that the IB index was higher in black and Hispanic subjects, those with less than a high school education, no alcohol intake, and without cardiac disease. Furthermore, the IB index was associated with greater odds of having MMSE ≤ 24 compared to MMSE > 24 (unadjusted odds ratio = 1.58). The effect of the IB index on MMSE did not differ by APOE genotype. The association between IB index and MMSE was prominent among subjects who were physically inactive, women, had Medicaid or no insurance, and had less than a high school education. However, no relation was found between IB and change in cognition over time based on the annual telephone interviews, either unadjusted or adjusted for demographics and risk factors (P = 0.13). Moreover, findings were similar with both the IB and VIB in that the VIB index was associated with MMSE ≤ 24 (adjusted odds ratio = 1.22; P = 0.04) but not with change over time during follow up interviews (P = 0.24).


There were several limitations to the study. While prior research supports a possible role for several of the infections that were tested (i.e., CMV, HSV-1), it is possible that other common viruses, like Epstein-Barr virus, Hepatitis B and C, can also lead to cognitive decline, although this remains theoretical. Furthermore, the authors did not test for HIV and its potential impact on the cohort is unknown. Because of the cross-sectional nature of the study, definitive conclusions about the direction of the associations (i.e., which came first, the infection or the dementia) cannot be determined with certainty. It was not possible to examine the relationship between infections and specific forms of cognitive impairment, such as AD and vascular dementia. Testing for syphilis, a known infectious etiology of dementia in late illness, was not performed. Finally, the role of threshold effect on the study (i.e., the damage is already done so there is no further decline) may have reduced the ability to detect an association between IB and cognitive decline over time.

The microbe-dementia hypothesis is intriguing because it implies that dementia could be reversible with antimicrobial agents. If true, it would signify a major paradigm shift in our understanding of cognitive decline. The study by Katan and colleagues extends previous findings of the association between chronic infections and cognitive decline. It is somewhat surprising that the IB index did not correlate with cognitive decline over time. As the authors hypothesized, this could be due to the relatively advanced stage of cognitive impairment at the time the subjects were enrolled, which would have limited the ability to detect further decline. The mechanism behind the association is uncertain but might be from the inflammatory response elicited by chronic infection. This inflammation, combined with other risk factors, then leads to atherosclerosis, subclinical stroke, and dementia. The close (essentially identical) correlation between the IB and VIB in the study supports the notion that most of the effect on cognition is mediated by viral rather than bacterial infections.

Although the study showed an association between IB and cognitive performance, association does not always equal causality. As noted in an accompanying editorial, only a randomized controlled trial would be definitive.2 Currently, there are no published clinical trials in humans on treating AD with antimicrobials. In a recent study, investigators administered minocycline to transgenic mice predisposed to AD-like amyloid pathology.3 They found the drug down-regulated inflammatory markers that correlated with a reduction in amyloid precursor protein levels and amyloid precursor protein-related products. Minocycline has known anti-inflammatory properties and most likely did not exert a direct antimicrobial effect. Following an initial animal trial, the next step could be a randomized controlled trial in patients with AD using valacyclovir over an extended duration, perhaps 6 months or longer, to treat reactivations of herpes viridae. Study endpoints would have to be chosen carefully and would likely require detailed and comprehensive neurocognitive testing. Anti-inflammatory therapy could also be given and APOE ε4 genotype determined. Such a study could produce solid evidence-based medicine that determines the validity of the microbedementia hypothesis.


1. Hurd MD, et al. N Engl J Med 2013;368:1326-1334.

2. Strandberg TE, et al. Neurology 2013;80:1182-1183.

3. Ferretti MT, et al. J Neuroinflammation 2012;9:62.