This systematic review has been registered with the National Medical Research Register, Ministry of Health Malaysia, and the protocol has been registered with PROSPERO (International Prospective Register of Systematic Reviews; CRD42024574930). The findings will be disseminated to relevant stakeholders involved in policymaking and program management for dengue surveillance. These results will also help in our future research on dengue wastewater surveillance to choose the best sampling and analysis methods. This review involves data extracted from previously published, publicly available sources and does not directly involve participants or the use of any identifiable personal data. Therefore, formal ethics approval was not required.

This protocol was prepared based on the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) checklist (Checklist 1). This review is still ongoing and the procedures outlined in this section will be applied during the review process. The systematic review and meta-analysis will be reported in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [22].

This review will include sewage samples from wastewater treatment plants serving population-based communities or public/private institutions.

We will include studies conducting WBS involving different sampling methodologies (sample type, sampling techniques, frequency, location), detection and serotype analysis.

There is no restriction of the included articles with or without a comparator.

The outcomes that will be included are viral load (gene copies, Ct value). The outcomes will be compared based on different contexts, which are sampling methodologies and detection methods (polymerase chain reaction–based, enzyme-linked immunosorbent assay) and geographical area.

In addition, we will include studies reporting serotype analysis, positivity rate (%), epidemiological trends (above or below the limits; stagnation, increase, or decrease in the number of cases), and molecular characterization of DENV.

There is no involvement of patients and the public in this review.

The eligibility criteria are as described in Textbox 1.

A comprehensive search will be performed using 4 databases (PubMed, Scopus, Web of Science, and Embase) by two authors independently based on the eligibility criteria. The search strategy will include a combination of predefined keywords that are Medical Subject Headings (MeSH) terms and free-text terms related to dengue WBS. Boolean operators such as OR and AND will be used to combine these keywords and refine the search findings in the titles and abstracts of databases. The search strategy to be used is presented in Checklist 1. Additionally, a manual screening of references from articles will be performed. The time frame for the inclusion of studies will be from inception until June 2025.

Articles retrieved from the electronic searches will be imported into Zotero software (version 6.0.37; Corporation for Digital Scholarship) to check for and remove duplicates. Subsequently, the records will be exported to a CSV file format. Unique identifiers will be given to the records and the data will subsequently be managed through Microsoft Excel (Microsoft Corp).

Preliminary title and abstract screening of the remaining articles will be done independently by two authors to retrieve relevant articles based on the predetermined criteria of the study population, intervention, comparator, and outcomes. Subsequently, full-text articles will be reviewed by two reviewers to assess their eligibility based on the exclusion and inclusion criteria using an eligibility form. If discordance to accept or reject an article arises, a third reviewer will be sought to come to a decision. Reasons for exclusion of articles will also be recorded and discussed. Two independent authors will identify articles to be included in the meta-analysis. The flow of the strategy used for the selection of the included articles will be summarized in a PRISMA diagram as a figure.

Comprehensive data extraction will be performed using a standardized form to systematically gather information from each study. This will include study characteristics (author details, publication year, country, study design), sampling methodology (sample type, sampling techniques, frequency, location), method of detection, and serotype analysis. To ensure the reliability of the data, two independent reviewers will perform the extraction, with any discrepancies resolved through discussion or consultation with a third reviewer. This will facilitate a thorough analysis of the evidence, emphasize significant findings, identify research gaps, and establish a basis for future studies in the field.

Quality and risk of bias will be appraised with design-specific instruments. The Risk of Bias in Non-randomized Studies – of Interventions (ROBINS-I V2) tool will be used to assess the quality of included studies. This tool evaluates the risk of bias across seven domains: (1) bias due to confounding, (2) bias in the classification of interventions, (3) bias in the selection of participants into the study (or into the analysis), (4) bias due to deviations from intended interventions, (5) bias due to missing data, (6) bias arising from measurement of the outcome, and (7) bias in the selection of the reported result [23]. Each domain will be rated as having a low, moderate, serious, or critical risk of bias [24]. Two independent reviewers will assess each study across these domains, and any discrepancies will be resolved by involving a third reviewer. The overall risk of bias for each study will be determined based on the “worst-score rule,” meaning that the final judgment reflects the highest level of bias identified in any single domain.

In addition to ROBINS-I, other tools may be considered based on study design. For cohort and case-control studies, the Newcastle-Ottawa Scale may be used, which judges study selection, comparability, outcome and exposure domains via a star system, and classifies overall quality (good/fair/poor) using established thresholds [25]. Diagnostic accuracy or method-comparison studies (eg, assay sensitivity and specificity against a reference standard) will be appraised with QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies tool), covering the case selection, index test, reference standard, and flow/timing that led to domain-level ratings (low, high, or unclear) and an overall judgment [26]. For analytical validation without a case-level reference, we will apply adapted QUADAS-2 signaling questions focused on sample handling, analyst blinding, repeatability, and selective reporting. In studies with overlapping designs or mixed components such as those reporting both exposure-outcome associations and diagnostic accuracy, we will apply more than one appraisal tool as appropriate to each component. In such cases, we will present risk-of-bias ratings separately for each tool and component. If this yields differing overall judgments, we will prioritize the tool corresponding to the study’s primary aim or default to the higher risk rating for synthesis purposes. The rationale for tool selection and integration will be documented. However, it should be noted that these alternative tools are generally less comprehensive than the ROBINS-I tool, particularly when adapted for assessing bias in exposure studies. A quality summary will be presented in a table or figure.

Two independent reviewers will complete domain-level ratings using prespecified guidance and standardized forms; disagreements will be resolved by discussion and, if necessary, a third reviewer, defaulting to the more high-risk rating when consensus is not achievable. We will document rationales for each judgment and present domain- and study-level summaries in tables or traffic-light plots.

Statistical analysis will be performed using Review Manager (RevMan; version 5.4; The Cochrane Collaboration). A meta-analysis will be conducted if at least two comparable studies with similar outcome measures are available. Meta-analyses will be conducted for viral load quantifications from included articles to determine overall effect sizes. Effect sizes will be calculated as risk ratios or odds ratios for categorical outcomes (eg, detection of DENV in wastewater) and as mean differences or standardized mean differences for continuous outcomes (eg, viral load or concentrations of dengue markers) [27]. Heterogeneity will be assessed using the I² statistic and Q test. The I² statistic will be used to quantify the proportion of total variation across studies attributable to heterogeneity rather than chance, with thresholds of 25%, 50%, and 75% representing low, moderate, and high heterogeneity, respectively [28]. The Q test, which will apply a χ² test, will be used to statistically evaluate heterogeneity among the studies.

Pooled analyses that use fixed-effects models for low heterogeneity (I²<50%) and random-effects models for moderate to high heterogeneity (I²≥50%), accounting for variability across studies, will be performed [29]. Subgroup analyses will be performed based on factors such as sampling techniques, detection methods, dengue serotypes, sampling sites, geographic location, and seasonality. Sensitivity analyses will be conducted to assess the robustness of the overall findings by excluding studies with high risk of bias or outlying results [30]. Cumulative meta-analysis will track trends in effect sizes over time [31]. Studies meeting inclusion criteria but lacking data for meta-analysis will be summarized narratively and compared with the meta-analysis findings.

The certainty of evidence will be systematically evaluated using the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) framework [32,33]. This evaluation will address (1) study limitations, considering potential biases in selection, performance, detection, and reporting that may affect the reliability of the results; (2) inconsistency, where variations in findings across studies may undermine confidence in the estimated effect size; (3) indirectness, which will assess the applicability of the evidence to the research question by examining differences in populations, interventions, and outcomes across studies; (4) imprecision, which will evaluate the certainty of the effect estimates, with wide confidence intervals reflecting uncertainty in the findings; and (5) publication bias [33]. Each of these domains will be rated to determine the overall quality of the evidence as high, moderate, low, or very low.

Publication bias will be assessed using a funnel plot by visualizing the effect sizes of different studies, particularly between small and large study sizes. Further tests including the Egger regression test significance at P<.05 [34] will be suggested if bias is suspected, given the number of articles included will be more than 10 [30].

This systematic review and meta-analysis will describe and synthesize the current evidence of dengue WBS in terms of the available methodologies and their correlation with viral concentration. The reviewed studies are expected to demonstrate the detection and quantification of DENV in population-based WBS and dengue-spiked experimental setups. Importantly, these studies will inform the selection of critical methodologies for conducting dengue WBS, including sampling strategies, sample types, concentration methods, nucleic acid extraction, and detection techniques. Additionally, they will offer insights into the feasibility and practical challenges of implementing WBS for dengue surveillance.

This will be the first comprehensive review on the global application of this surveillance method for dengue. WBS has been increasingly recognized as a cost effective and non-invasive tool to monitor infectious diseases in the community setting and its application has great potential to enhance public health measures and improve disease outcomes [36]. The data from the WBS could complement clinical surveillance to predict possible dengue outbreaks and provide an early warning through the detection of asymptomatic cases. This review will also guide the selection of key methodologies in performing dengue WBS and identify the gaps and limitations of different sampling and detection methods.

Given the emerging status of this field, we anticipate a limited number of studies will be available, which could limit the comparability and meta-analysis of the sampling and detection methods. There is also a possibility of reporting biases, with a higher frequency of positive results being published than negative findings.