Methods and Applications: Wastewater-Based Monitoring of Dengue Fever at Community Level — Guangzhou City, Guangdong Province, China, May 2024

Serum and urine samples from confirmed cases (Supplementary Material) were obtained from hospitals or district CDCs in Guangzhou. On the day of diagnosis, 1 mL of serum was collected from each case. Subsequently, a total of 36 midstream urine samples (50 mL each) were collected from three cases. Given the urgent need for epidemic response, grab wastewater samples were collected as a practical alternative to 24-hour composite sampling. In total, 618 grab samples were obtained from wastewater manholes within 200 m of all confirmed case residences — a distance corresponding to the typical activity range of Aedes mosquitoes (Figure 1). Additionally, 24-hour composite influent samples were collected from WWTPs to enable a comprehensive assessment of detection sensitivity. Each wastewater sample volume was 500 mL. Wastewater surveillance encompassed all of Guangzhou during the initial wave of local dengue fever outbreaks, with sampling conducted at 115 distinct sites. All samples were transported at 4 °C to the laboratory for analysis within 24 hours.

Figure 1. 

A schematic layout of sampling sites from wastewater manholes within 200 m of the residence of the first locally acquired dengue case (Case 1).

Serum RNA was extracted using the magnetic bead virus nucleic acid extraction kit (Jiangsu Bioperfectus Technologies Co., Ltd, Taizhou, Jiangsu, China). A 2 mL-aliquot of the urine sample was used for one-step virus concentration and RNA extraction using a magnetic bead-based biological sample virus concentration kit (Suzhou Advanced Molecular Diagnostics Co., Ltd, Suzhou, Jiangsu, China).

Two concentration methods, polyethylene glycol precipitation and the magnetic bead-based method, were systematically compared in terms of limit of detection (LOD), PCR inhibition, and recovery efficiency (Supplementary Material). Primers and probes used are shown in Supplementary Table S1. The method demonstrating superior performance was selected for wastewater sample concentration.

To further confirm the type of DENV infection in patients, reverse transcription PCR (RT-PCR) of serum RNA was conducted on four serotypes. The RNA concentration of DENV-1 in wastewater and urine samples was obtained by RT-qPCR. Sanger and whole-genome sequencing of DENV-1 was performed in positive serum, urine, and wastewater samples. Sequence analysis and visualization were carried out. The detailed descriptions are shown in the Supplementary Material.

For DENV-1 detection in wastewater samples, spiked murine hepatitis virus (MHV) served as the sample processing control (9). Following RNA extraction, the OneStep^TM PCR Inhibitor Removal Kit (Zymo Research, CA, USA) was employed to eliminate potential inhibitors.

DENV-1 target sequences were successfully obtained from serum samples of 6 out of 8 cases, with GenBank (National Center for Biotechnology Information genetic sequence database) accession numbers PQ326420-PQ326425. Phylogenetic tree analysis revealed that the DENV infections in 2 cases belonged to DENV-1 genotype I (branch 1 and branch 2), while the DENV infections in 4 cases belonged to DENV-1 genotype III (branch 3) (Figure 2). Through online Nucleotide Basic Local Alignment Search Tool (BLAST) analysis, branches 1, 2, and 3 showed the closest similarity to KY057370 (isolated in 2012, Indonesia), OQ678061 (isolated in 2019, Cambodia), and MZ312929 (isolated in 2018, India), respectively. These sequences demonstrated close phylogenetic relationships with Southeast Asian strains, consistent with genomic analysis results from previous dengue cases in Guangzhou (10).

Figure 2. 

The phylogenetic tree of dengue virus serotype 1 (DENV-1) branches in serum samples.

Urine samples demonstrated a 100% positivity rate for DENV-1, with RNA concentrations ranging from 7.0 to 50,065.5 copies/mL, consistent with previously reported values (10² to 10⁴ copies/mL) (6). Prior studies (6,11–12) indicate that the primary source of detectable DENV RNA in wastewater originates from patient urine, whereas SARS-CoV-2 RNA predominantly derives from fecal shedding, with reported concentrations ranging from 10² to 10⁸ copies/mL (13). Consequently, the substantially lower concentration of DENV in wastewater compared to SARS-CoV-2 presents a significant analytical challenge for wastewater-based surveillance of dengue fever.

Both concentration methods demonstrated a LOD of 10 copies/mL (Supplementary Table S2), with no evidence of PCR inhibition (Supplementary Table S3). However, the magnetic bead-based method exhibited superior recovery efficiency (59.7%) compared to polyethylene glycol precipitation (50.3%) (Supplementary Table S4). Given its enhanced recovery rate and automation potential, the magnetic bead-based method was selected for routine wastewater monitoring. In total, 14 wastewater samples tested positive for DENV-1, yielding a positivity rate of 2.3%. All positive samples were collected from wastewater manholes within 50 m (straight-line distance) downstream of patients’ residences during the initial 5 days of sampling. Sanger sequencing results of the RT-qPCR products from positive wastewater samples further confirmed the detection of DENV-1 sequences. The sequence alignment coverage reached 100.0%, with similarity ranging from 98.2% to 100.0%. RNA concentrations ranged from 0.7 to 97.5 copies/mL. These values were slightly lower than those reported in WWTPs (31–450 copies/mL) (8), highlighting potential differences in sampling environments and viral persistence.

Whole-genome sequencing revealed that the read counts of the obtained DENV-1 sequences from wastewater, urine, and serum samples from the same patient were 6,311/559,029 (1.1%), 206,362/341,880 (60.4%), and 1,573,040/1,590,306 (98.9%), respectively (Figure 3). These results indicate that the degradation of DENV-1 sequences in wastewater is substantially greater than in serum and urine samples. This finding suggests that wastewater-based monitoring and whole-genome sequencing of dengue fever present greater technical challenges compared to clinical specimens.

Figure 3. 

Distribution of dengue virus serotype 1 (DENV-1) sequences in serum, urine, and wastewater samples based on whole-genome sequencing. (A) The distribution and coverage depth of all sequences; (B) The distribution of sequences with a coverage depth of no less than 10.

Ongoing dengue control efforts have successfully maintained low case numbers in Guangzhou. Between 2010 and 2019, annual dengue incidence fluctuated from 58 to 1,925 cases, with a significant outbreak in 2014 resulting in 38,036 reported cases (10). In recent years, dengue cases have remained consistently low, with only 9–37 cases reported annually from 2020 to 2022. Dengue virus RNA concentrations in urine, as documented in this and previous studies, range from 10¹ to 10⁴ copies/mL (6). Assuming an average daily urine output of 2 L per person, daily dengue virus shedding is estimated at 2×10⁴ to 2×10⁷ copies per individual, approximately 100 times lower than SARS-CoV-2 shedding [1.27×10⁶ to 1.04×10⁸ copies/(day·person)] (12). The combination of low dengue prevalence and substantially reduced viral shedding creates considerable challenges for wastewater-based surveillance. Unlike SARS-CoV-2, which remains detectable in wastewater from large catchments even with minimal case numbers, dengue virus was only detectable within a 50 m radius of infected residences, with WWTP samples falling below detection limits. Under these conditions of low case numbers and limited viral shedding, wastewater-based surveillance at WWTPs provides insufficient sensitivity for effective outbreak detection. Instead, robust dengue surveillance requires targeted sampling at the community level, with strategic focus on locations proximal to identified cases.

Despite these challenges, our findings demonstrate the significant potential of wastewater-based surveillance as a complementary tool for community-level dengue monitoring. Notably, all positive wastewater samples in this study correlated precisely with epidemiological survey results, with no false-positive detections observed. More importantly, wastewater surveillance demonstrates promise for early detection of asymptomatic or pre-symptomatic infections, enabling timely interventions. The first locally acquired dengue case was clinically diagnosed and hospitalized at 18:00 on May 2. Later that evening at 23:00, wastewater monitoring and epidemiological surveys were initiated (Figure 4). Remarkably, despite the hospitalization of this initial case, positive wastewater signals persisted at Site 1 and Site 2 from May 2 to May 6. This unexpected persistence prompted further investigation. Since grab wastewater samples were collected from manholes beneath the patient’s building rather than from wastewater storage tanks, we excluded the possibility that the detected viral signals originated from Case 1. Subsequent epidemiological investigations identified a second infected individual residing in the same building at 19:00 on May 3. This patient remained at home under observation for three days before hospitalization on May 6. The spatiotemporal correlation between wastewater viral concentrations and confirmed cases strongly indicates that the wastewater signals at Site 1 and Site 2 were attributable to Case 2. These findings underscore the utility of wastewater surveillance in identifying undetected transmission chains. Furthermore, trends in wastewater viral levels provide valuable insights into the temporal dynamics of outbreaks, offering a complementary tool for real-time epidemiological monitoring and early intervention strategies.

Figure 4. 

Schematic diagram of wastewater surveillance identifying Case 2 prior to clinical diagnosis. (A) Timeline of clinical confirmation/hospitalization for Case 1 and Case 2, and corresponding positive wastewater signals. (B) Relative geographical locations of Case 1 and Case 2 residences, and case-origin attribution of positive wastewater samples.

Abbreviation: DENV-1=dengue virus serotype 1.

Although wastewater monitoring for dengue frequently yields negative results, these findings provide valuable insights for tracking epidemic trends. For instance, wastewater surveillance continued for nine days following the hospitalization of Case 2, during which all wastewater samples collected near the residence tested negative — consistent with the absence of further infections. A similar pattern was observed for other patients, as no additional cases emerged during their hospitalization, aligning with the lack of detectable DENV-1 in wastewater samples collected around their residences. These results highlight the potential of wastewater surveillance as a valuable tool for assessing the ongoing risk of dengue transmission and determining whether continued prevention and control measures are necessary in a given area.

While this study successfully demonstrated the potential of wastewater-based surveillance for early detection and outbreak monitoring, several limitations warrant consideration in future research. First, the study relied on grab sampling, which may have reduced detection sensitivity. Given the generally lower concentrations of DENV compared to SARS-CoV-2, employing larger-volume samples, 24-hours composite sampling, or solid-phase extraction methods (which enhance concentration factors) could improve detection sensitivity (3,8,14–15). Enhanced sensitivity would extend the effective monitoring range, increasing the likelihood of detecting infections at a given density of sampling sites. Second, sequence analysis of DENV-1 ribonucleic acid (RNA) from urine and wastewater samples revealed significant viral degradation in wastewater. As samples were collected in close proximity to infected residences, degradation is unlikely to have occurred within the sewer system. Instead, it is more plausible that RNA degradation resulted from suboptimal conditions during sample transport and storage prior to analysis. Future studies should focus on optimizing these processes to minimize degradation, potentially through the use of stabilizing agents or rapid processing techniques to preserve viral RNA integrity. Third, this study was conducted during a period of sporadic dengue cases in Guangzhou, whereas dengue outbreaks typically peak in October and November. Ongoing surveillance efforts aim to assess the correlation between wastewater detection rates at WWTP inlets and the number of clinically diagnosed cases within corresponding catchment areas. This extended monitoring will provide critical insights into the feasibility of wastewater-based surveillance for outbreak prediction at a larger scale.