Association between herpes simplex virus 1 and dementia: a systematic review protocol

Abstract

Introduction

Herpes simplex virus 1 (HSV-1) infects approximately two-thirds of the global population under the age of 50 years. Although widely prevalent, the possible implications of HSV-1 in neurodegenerative diseases, especially dementia and Alzheimer’s disease, remain poorly understood. This review seeks to elucidate this association and explore the potential benefits of preventing or treating herpesvirus infections on dementia risk. The goal is to enhance our understanding of HSV-1’s potential role in dementia, which could inform the development of future therapeutic interventions for these conditions.

Methods and analysis

PubMed, Embase (Elsevier/Ovid), Web of Science, Scopus, Global Health, PsycInfo, Cochrane Library and Clinicaltrials.gov will be searched from the inception of each respective database. Studies that have HSV-1 as an exposure and dementia, or its subtypes, as a primary outcome will be included. Two researchers will independently screen titles, abstracts and full texts, with discrepancies resolved by a third researcher. Systematic data extraction from eligible studies will be performed using a standardised template. Risk of bias of individual studies will be assessed with the Cochrane Collaboration approach. We will assess the overall quality of cumulative evidence using the Grading of Recommendations, Assessment, Development and Evaluations criteria. Statistical analysis will employ a random effects model, and heterogeneity will be determined with Cochrane’s Q test and assessed using I². Studies will be grouped by population subgroups and dementia subtypes when possible to explore nuances in results. We will consider performing meta-regression if heterogeneity remains after subgroup analyses. All statistical analyses will be conducted using Stata V.18 software (College Station, Texas, USA).

Ethics and dissemination

No ethical approval is required since data will be collected from existing studies. The review will be disseminated through peer-reviewed publication and at national and international conferences.

PROSPERO registration number

CRD42024516789.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This systematic review focuses specifically on herpes simplex virus 1 (HSV-1), the most prevalent herpesvirus in the world, to understand its nuanced effects on dementia development.

• The study only includes articles published in English.

Introduction

Dementia-related deaths are the sixth leading cause of death in the USA.¹ The prevalence is expected to increase as the population ages, with the number of dementia cases expected to reach 152.8 million globally by 2050 from the estimated 57.4 million cases as of 2019.² Alzheimer’s disease (AD) is the leading cause of dementia and is associated with over 80% of dementia diagnoses.³ Given that only one-third of AD cases may be attributed to modifiable risk factors and the lack of effective treatments for AD, much research has gone into uncovering other aetiologies and prevention methods to reduce AD and dementia risk.⁴ Specific causes of AD and dementia have been postulated and vigorously studied, with increasing evidence from large population-based studies citing viral aetiologies of dementia and cognitive impairment.⁵,⁷

The potential viral causes of dementia, including herpes simplex virus 1 (HSV-1), have been a subject of speculation for decades.⁸,¹⁰ HSV-1 is a neuroinvasive and neurotoxic virus capable of entering the brain via the peripheral nerves, and it is thus a candidate pathogen for increasing dementia risk.^{11 12} In 2016, an estimated 67% of the world’s population from the ages of 0 to 49 years had infection with HSV-1.¹³ Its prevalence and highly communicable oral–oral transmission lend it a critical factor when considering its role in neurological diseases.^{11 14} Randomised controlled treatment trials are currently underway to investigate the role of antiviral treatment of HSV-1 in AD.^{15 16} However, evidence at the population level remains mixed, and currently, there are no established antiviral strategies for reducing the risk of dementia.⁷

Previous systematic reviews have examined the link between herpesviruses and dementia but reported inconsistent findings and low-quality evidence, especially for HSV-1. A comprehensive review in 2019 analysed 32 case–control and 3 cohort studies on HSV-1’s impact on dementia, finding no significant difference in HSV-1 detection in the brains of dementia patients compared with controls (pooled OR 1.31; 95% CI, 0.90 to 1.90).¹⁷ In contrast, a more recent meta-analysis from 2023 identified a significant association between HSV infection and dementia incidence.¹⁸ However, it did not differentiate between HSV-1 and HSV-2, limiting its insights into the specific impact of HSV-1 on AD. Given HSV-1’s higher prevalence in the general population and its established link to AD in both epidemiological and brain studies, failing to differentiate HSV-1 from HSV-2 may obscure the specific risk associated with HSV-1.¹⁹,²³ Although HSV-2 has been broadly associated with AD in brain studies, the relationship between HSV-2 infection and dementia in population studies remains inconclusive.^{18 24 25} Therefore, the lack of differentiation in the 2023 meta-analysis may obscure the more pertinent risk posed by HSV-1, highlighting the need for research focused specifically on HSV-1 to clarify its association with dementia.

Given these conflicting findings, our study seeks to update existing reviews with a specific focus on HSV-1. The present report presents the meta-analysis protocol with results from this study anticipated in June 2025. By incorporating recent studies and leveraging subtype analyses and meta-regression, we aim to explore how HSV-1 impacts dementia risk across demographics such as age, ethnicity, socioeconomic status and immune status. This approach not only refines the current understanding but also informs the design of future randomised controlled trials (RCTs) targeting HSV-1 prevention as a potential strategy to reduce dementia risk.

Methods

This systematic review protocol has been prepared according to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) Protocols 2020 guidelines.²⁶ The data collectoin began in February 2024 and analyses concluded in March 2025.

Data organisation

Search results will be uploaded into an EndNote software (V.21) and deduplicated.²⁷ Study selection will be managed through Covidence. Presentation of results will align with guidelines in the Cochrane Handbook for Systematic Reviews of Interventions²⁸ and the PRISMA statement.²⁶

Eligibility criteria

Inclusion

Types of Studies

Studies must be written in English to be considered.

The following studies will be included: randomised control trials; observational studies, including cross-sectional, prospective and retrospective cohorts, and case–control studies that present an estimate of effect or provide sufficient evidence for an effect estimate to be calculated; brain specimen-based studies of deceased patients with a comparison group.

Population demographics

Studies with participants aged 18 years or older will be included. When possible, results will be stratified by age group, apolipoprotein E gene (APOE)-4 status, immune status and prior treatment history. All study settings will be considered and can be drawn from any healthcare or community setting, including but not limited to inpatient hospital care, outpatient services and primary care environments.

Exposure

Exposures will be infection with or reactivation of human HSV-1 infections (defined clinically, through appropriate laboratory criteria, or imaging in brain specimens). Treatments for herpes virus will also be considered (eg, with antiherpetic agents such as acyclovir, valacyclovir, etc).

Comparators

The control or comparison group will depend on the specific study design. Those might include: individuals not infected with HSV-1 in randomized-controlled trials and observational studies, or time periods before HSV-1 exposure in self-controlled case series or cross-over studies.

Outcomes

The primary outcome is the incidence of dementia, in any form, as confirmed through clinical diagnosis, neuroimaging or histopathology findings. Where data permit, we will further classify dementia types, such as AD, vascular dementia and frontotemporal dementia, to understand the potential differential effects of HSV-1 infection.²⁹

Exclusion

Studies will be excluded if they are animal studies, grey literature, do not differentiate between HSV-1 and HSV-2, or do not present binary outcome measures (eg, cognition reported as a linear scale rather than the presence or absence of cognitive impairment). Additional reasons for exclusion include the absence of a comparator group or insufficient information to calculate an effect estimate (eg, HSV-1 infection coupled with another viral infection).

Search strategy

Studies will be systematically searched through the following databases:

PubMed, Embase (Elsevier/Ovid), Web of Science, Scopus, Global Health, PsycInfo, Cochrane Library and Clinicaltrials.gov will be searched after inception of each respective database for efficient coverage of the subject matter.³⁰ The search terms included medical subject header terms and free text keywords. The basic structure of the search strategy included search terms related to (‘herpes simplex virus 1’) AND (‘dementia’ OR ‘Alzheimer’s disease’). These terms will be searched within the title, abstract, keywords and the main body of the articles. The provisional search terms are listed in the online supplemental appendix 1.

Selection process

After a preliminary search in February 2024, 3406 articles were identified from eight databases (online supplemental appendix 1). The search is planned to be updated prior to completion of the final review to ensure that it contains the most accurate and recent data. Two researchers will independently screen the titles and abstracts (FL and APV), and discrepancies will be resolved with a third reviewer (KAL). Subsequently, the full texts of articles that are screened in will be acquired and reviewed independently by the same researchers, and discrepancies will be resolved with a third reviewer. Reasons for exclusion will be documented. The record review and selection process will be illustrated using a PRISMA flow chart.²⁶

Data extraction

Data extraction will be performed by two independent reviewers (FL and APV) using a standardised extraction table. The extraction data will include specific details organised across several categories. For study characteristics, we will collect information on authors, publication year, study design, period of study and length of follow-up time. Regarding the study population, the data will cover case and control characteristics such as sex, age, ethnicity, socioeconomic status, APOE4 status and antiviral therapy status, alongside recruitment and sampling methods and the duration between HSV-1 infection and dementia diagnosis, if available. Exposures will be documented by classifying HSV-1 status and its detection method, as well as counting the number of HSV-1-infected subjects. Comparators will involve the classification and identification of HSV-1-uninfected individuals and the number of HSV-1-uninfected subjects. Outcomes will be categorised by the classification and diagnostic criteria of primary outcomes (dementia) and secondary outcomes (dementia subtypes), and the number of subjects with the outcome will be recorded when available. Finally, effect measures will include effect estimates such as risk ratios (RRs), hazard ratios (HRs), odds ratios (ORs), probability (p) values and confidence intervals (CIs), as well as adjusted effect estimates.

Covariates measured and adjusted for in each study will be recorded and included in subsequent analysis, if indicated. If there are multiple publications of one study, we will count it as one study, primarily referencing the main report and extracting additional data from secondary studies. Authors of published studies will be contacted with a standardised email template, in the cases where additional information is required, such as methods or results for data analysis.

Quality assessment

We will evaluate the overall quality of evidence from multiple studies regarding the association between HSV-1 and each outcome. This assessment will employ specific criteria from the Grading of Recommendations, Assessment, Development and Evaluations framework, tailored to the characteristics of the included studies.³¹ We will assess for risk of bias, inconsistency, indirectness, imprecision, publication bias, magnitude of effect, dose–response gradient and any other relevant domains to classify the strength of evidence as ‘high’, ‘moderate’, ‘low’, ‘very low’ or ‘unclear’.

Assessment of the risk of bias in included studies

Bias risk will be evaluated in individual studies qualitatively following the Cochrane Collaboration approach, applying it to both randomised and non-randomised studies.³²,³⁴ For randomized-controlled trials, this will include consideration of the following domains: random sequence generation and allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias) and other biases. For observational studies, we will consider preinterventional domains (selection of participants and control for confounding variables), interventional domains (classification of interventions) and postinterventional domains (missing data, measurement of outcomes and selective reporting of results). For each study, each component will be assigned a risk of bias category ‘high risk’, ‘moderate risk’, ‘low risk’ or ‘unclear risk’, as defined in the Cochrane handbook. A summary risk of bias will be presented in table format. Two reviewers will assess risk of bias categories, with any unresolved discrepancies resolved through discussion with a third reviewer.

Outcome measures

Our primary focus is dementia, including its subtypes, which we will identify and document wherever possible. In evaluating the quality of studies, our emphasis will be on those that verify infection with HSV-1 before the onset of dementia-related outcomes, as well as studies that provide a clear timeline from HSV-1 infection to diagnosis of dementia. We will also present any evidence regarding the potential impact of antiviral treatment against HSV-1 on dementia.

While our main interest lies in longitudinal prospective studies, we are also open to considering the ORs provided by case–control studies for a comprehensive analysis. Adjusted effect estimates and methods of adjustment will be calculated for the association between HSV-1 and risk or rate of dementia. Results of any additional stratified analyses will also be recorded, for example, on virus effects in different age groups or on dementia subtypes, including categories of severity.

Given that cognitive impairment exists on the spectrum of dementia but may not meet the clinical definition of dementia, we will conduct a sensitivity analysis including studies that report on cognitive impairment without a specified dementia diagnosis, as well as those that consider both cognitive impairment and dementia together. This approach balances specificity in our primary outcome while ensuring we have sufficient statistical power to detect potential effects.

Data synthesis

All statistical analyses will be conducted using Stata V.18 software (College Station, Texas, USA).

When applicable, ORs will be converted to RRs. In sensitivity analysis, studies with multiple domains classified as high risk of bias will be excluded to assess their impact. Additionally, leave-one-out cross-validation will be conducted to evaluate the validity and robustness of our meta-analyses.

We will use a random effects model to obtain a pooled RR across all studies. A forest plot will be generated to summarise all results. If there is significant heterogeneity, we will perform subgroup analyses according to study design, year of publication or study quality. If necessary, we may perform a multivariable meta-regression if there are an adequate number of studies that can be categorised based on dementia subtypes or population subtypes to compute pooled effect estimates.

Assessment of heterogeneity

To assess the presence of heterogeneity, we will use Cochrane’s Q test in detecting heterogeneity; a p value≤0.05 will indicate significant heterogeneity and additional stratified sub-analyses will be presented.³⁵ Additionally, I² will also be assessed and, if significant, would also trigger additional analyses.³⁶ Galbraith plots will be generated to further investigate heterogeneity and to detect potential outliers. If many studies are identified for inclusion, this plot will be presented to summarise the pooled RR. L'Abbé plots will be generated to compare binary outcomes between groups to further explore potential heterogeneity by comparing study-specific event rates in the two groups.

Small study effects and publication bias

To explore the possibility of small study effects and publication bias, we will generate funnel plots and test for small study effects, including moderators as needed to account for potential between-study heterogeneity.³⁷ Last, we will perform a trim-and-fill analysis to account for publication bias, which estimates the number of studies potentially missing from the meta-analysis due to publication bias.³⁸ If applicable, Egger’s regression test and Begg’s rank correlation test will be used to assess the asymmetry of funnel plots and potential publication bias in the meta-analyses, particularly if fewer than 10 studies are included in each analysis.^{39 40} Potentially missing studies are imputed, and the pooled RR is estimated using both observed and imputed studies.

Patient and public involvement

Members of the public were not involved in the design or implementation of the study but were included in the initial dissemination of the study protocol. Methods of the study were presented and reviewed at a monthly meeting with Herpes Cure Advocacy, a non-profit, patient-led advocacy group. Results will be shared with Herpes Cure Advocacy on completion of the study.

Ethics and dissemination

As this is a review of existing studies, no ethical approval is required. Any changes to the protocol will be published on PROSPERO under registration number CRD42024516789. Our findings will be disseminated through a peer-reviewed publication and at national and international conferences. We anticipate the review will clarify the existing scope and reliability of evidence for a link between human herpes simplex virus 1 and dementia, identifying gaps and informing the direction of future therapies and research.

The views expressed in this publication are those of the authors and do not reflect the official position of Keck School of Medicine of the University of Southern California.