Spatial patterns of dengue cases and Wolbachia Occurrence in Aedes mosquitoes in a persistent endemic area of Samarinda, East Kalimantan, Indonesia

Persistent dengue endemicity remains a major public health concern in tropical and subtropical regions. Spatial analysis offers critical insights into transmission dynamics and hotspot distribution, supporting more targeted interventions. This study examined persistent dengue transmission in a hyperendemic setting by evaluating vertical transmission, Wolbachia establishment, and spatial heterogeneity. The epidemiological analysis included all 1,081 laboratory-confirmed dengue cases reported between January and December 2024 across 59 villages in Samarinda, East Kalimantan, Indonesia. A nested entomological survey included all 150 eligible households of laboratory-confirmed dengue cases identified during field surveillance between June and September 2024 across 40 villages, using 300 indoor and outdoor ovitraps. Female Aedes aegypti reared from field-collected eggs and egg samples were tested for DENV, while female Aedes aegypti were screened for Wolbachia. Spatial patterns were evaluated using Global Moran’s I, Local Moran’s I, and Getis-Ord Gi statistics. Forty of 59 villages (67.8%) were classified as highly endemic. DENV was not detected in the adult mosquito and egg samples tested. Wolbachia was detected at 39.9% of trapping locations with valid molecular results. Dengue incidence showed significant spatial clustering, whereas Wolbachia occurrence was spatially heterogeneous and overlapped with some dengue hotspots. Failure to detect DENV in the tested samples does not exclude rare vertical transmission. Likewise, the spatial co-occurrence of Wolbachia and dengue hotspots cannot establish the origin or epidemiological effectiveness of Wolbachia. Nevertheless, the integrated surveillance approach provides a useful framework for targeted, context-specific dengue surveillance and vector-control planning.

Dengue remains a critical global public health concern, with rising incidence reported in numerous countries in recent years. It is estimated that approximately 100–400 million new dengue infections occur annually, predominantly affecting developing nations (1, 2). In alignment with the global target of achieving zero dengue deaths, the Indonesian Ministry of Health has established a national strategy aiming to reduce the incidence rate to 49 per 100,000 population by 2025, with a long-term goal of zero deaths by 2030 (3).

Despite these efforts, national reports indicate a fluctuating trend in dengue cases over the past decade. In 2023, the national Incidence Rate (IR) was recorded at 41.36 per 100,000 population with a Case Fatality Rate (CFR) of 0.73% (4). However, the situation varies significantly at the local level. Samarinda, the capital of East Kalimantan province, represents a city with consistently high endemicity over the last 12 years. In 2023, Samarinda reported an IR of 203.1 per 100,000, which is nearly five times the national average, and a CFR of 0.53% (5). This study indicates that the distribution of cases in this metropolitan area follows a clustered pattern, often grouped within a 100-meter radius of high-risk households.

This persistent high dengue endemicity is driven by multiple factors, including the co-circulation of multiple serotypes, viral mutations, vertical transmission (VT) within mosquito populations, and a high proportion of asymptomatic infections (6–9). These biological mechanisms are further reinforced by high vector density, rapid urbanization, and the limited effectiveness of conventional control strategies (10). Recently, Wolbachia -based interventions have emerged as a promising innovation. East Kalimantan Province has become a focal point for this strategy, where Bontang City, located approximately 120 km north of Samarinda, has served as a pilot site for the artificial introduction of Wolbachia -infected Aedes aegypti mosquitoes (11). This proximity creates a unique epidemiological situation regarding potential spread of Wolbachia in neighboring urban centers like Samarinda, although its long-term impact on transmission dynamics in these hyperendemic settings needs to be clarified (12, 13).

Hence, to address these challenges, we need effective, safe, and sustainable vector management as a key strategy (14–16). The utilization of spatial analysis to map entomological predictors, genetic diversity, and human mobility is essential for generating accurate risk maps (17–20). Furthermore, understanding the mechanism of viral maintenance is crucial. Although transovarial (vertical) transmission has been documented in several regions of Indonesia (21–25), its occurrence in Samarinda remains unverified. This lack of data limits the current understanding of persistent endemicity in the region.

Therefore, this study aimed to investigate the dynamics of persistent dengue transmission in a hyperendemic setting. Specifically, we examined the role of vertical transmission, the establishment of Wolbachia in local mosquito populations, and the spatial heterogeneity of both the virus and the vector to inform more targeted control interventions.

Study design and area

This study employed a cross-sectional design and was conducted from June to September 2024 in Samarinda City, East Kalimantan Province, Indonesia. The study population consisted of dengue cases that were confirmed by laboratory NS1 antigen tests and reported to the Samarinda City Health Office. These cases were distributed across 40 administrative villages. The location of each confirmed case was geocoded to establish precise geographic coordinates for spatial analysis.

Entomological sample collection

Entomological surveillance followed WHO guidance (26). A pair of ovitraps, one indoors and one outdoors, was deployed for each of the 150 index households, yielding 300 traps. Traps were positioned in the immediate peridomestic area within approximately 35–50 m of the index residence and were monitored for seven days. This focal radius was selected to characterize mosquito activity around the index household and is consistent with the predominantly short-range, household-associated dispersal of Ae. aegypti described in field studies (27). Ovitrap contents were transported to the Microbiology Laboratory, Faculty of Medicine, Mulawarman University. Eggs were counted and reared, and emerged adults were identified morphologically. A trapping location was classified as Wolbachia -positive if at least one mosquito pool collected from that location tested positive. Female Ae. aegypti were selected for molecular testing.

Epidemiological data and classification

Secondary data regarding dengue patients in Samarinda throughout 2024 was obtained from the routine surveillance system of the Samarinda City Health Office to serve as a study reference. This dataset included the total number of laboratory-confirmed dengue cases and the population data for all administrative villages (n=59) in the city. The epidemiological status of each village was quantified by calculating the Incidence Rate (IR), expressed as the number of cases per 100,000 population. Villages were classified based on the criteria from the Indonesian Ministry of Health: High Endemic (IR ≥ 49 per 100,000), Endemic (IR < 49 per 100,000), or Not Endemic (zero cases). This data served as the foundation for analyzing the spatial heterogeneity of dengue incidence across the study area.

Laboratory procedures

Field-collected Aedes eggs were reared to adulthood under insectary conditions, and only emerged adult female Ae. aegypti were included in the molecular analyses; eggs were not tested directly. A total of 631 adult female Ae. aegypti were analysed in pools of 10 mosquitoes per pool. For DENV detection, total RNA was extracted using the Zymo Quick-RNA Viral Kit (Zymo Research, USA) according to the manufacturer’s instructions, and RNA concentration and purity were assessed using a NanoDrop spectrophotometer. DENV detection and serotyping were performed by conventional reverse-transcription PCR (RT-PCR) using the consensus primers D1/D2 and the serotype-specific primers TS-1, TS-2, TS-3, and TS-4. For Wolbachia detection, genomic DNA was prepared from mosquito pools using the boiling extraction method and amplified by conventional PCR targeting the wsp, 16S rRNA, and ftsZ genes. PCR amplification was performed using an Applied Biosystems GeneAmp PCR System 9700 (Applied Biosystems, Foster City, CA, USA), and amplification products were resolved by agarose gel electrophoresis. Positive controls, extraction-negative controls, and no-template controls were included in each PCR run. PCR results were considered valid only when the expected amplicon was detected in the positive control and no amplification was observed in the negative controls. A Ct value between 1 and 35 indicated a positive result for Wolbachia, whereas values of 0 or greater than 35 were considered negative (28).

Statistical analysis

Statistical methods were employed to analyze the spatial distribution of dengue cases and the presence of Wolbachia. To quantify the overall spatial autocorrelation, Global Moran’s I was used to determine whether dengue cases were clustered, dispersed, or randomly distributed across the study area (29). To identify spatial heterogeneity at the local level, Local Moran’s I was applied. Spatial autocorrelation of village-level dengue incidence rates and the Wolbachia index was assessed using Global Moran’s I. Local Indicators of Spatial Association (LISA) were used to identify significant spatial clusters and spatial outliers, classified as high-high (HH), low-low (LL), high-low (HL), and low-high (LH). Hotspot and cold spot analyses were performed using the Getis-Ord Gi statistic* (30, 31). Spatial analyses were conducted using ArcGIS Desktop version 10.8 (Esri, Redlands, CA, USA). Spatial relationships were defined using a first-order Queen contiguity spatial weights matrix derived from Thiessen polygons and row-standardized before analysis. Statistical significance was assessed using 999 random permutations, with p < 0.01 considered statistically significant. Because the spatial analyses were exploratory, no adjustment for multiple testing was applied. The spatial analyses were descriptive and ecological and were not intended to infer a causal relationship between Wolbachia occurrence and dengue incidence.

Descriptive epidemiology

From 1 January through 31 December 2024, 1,081 laboratory-confirmed dengue cases were reported among a population of 856,360, corresponding to an IR of 126.23 per 100,000 population (Table 1). Of the 59 villages, 40 (67.80%) were classified as high endemic, 16 (27.12%) as endemic, and 3 (5.08%) reported no cases. Thus, 56 of 59 villages (94.92%) reported at least one dengue case during 2024.

Figure 1a shows the distribution of village-level case counts. The largest counts occurred in northern and central villages, including Sempaja Utara, Sempaja Timur, Lempake, Gunung Kelua, and Air Hitam; selected villages recorded more than 30 cases, with the maximum reaching 91. Lower counts were recorded in several peripheral and southern villages, including Bantuas, Handil Bakti, and Tani Aman. Figure 1b shows the endemicity classification. High-endemic villages were distributed across much of northern and central Samarinda, whereas Makroman and Rawa Makmur were classified as endemic and Gunung Panjang, Sungai Dama, and Budaya Pampang reported no cases. Bugis had a high IR relative to its population denominator.

The spatial distribution of dengue cases is shown in Figure 3. Figure 3a demonstrates a clear concentration of cases in the northern and central urban areas of Samarinda, particularly in Sempaja Utara, Sempaja Timur, Lempake, Gunung Kelua, and Air Hitam, where case counts exceeded 30 and reached up to 91. In contrast, peripheral and southern villages generally reported fewer than three cases. Figure 3b shows that most northern and central villages were classified as High Endemic (HE), whereas only a few villages were categorized as Endemic (E) or Non-Endemic (NE).

Entomological and virological findings

Entomological surveillance was conducted at 150 NS1-positive index households between June and September 2024. A total of 300 ovitraps were deployed, and all were successfully recovered. Eighty-five ovitrap paddles (85/300, 28.3%) were positive for mosquito eggs. Eggs collected from the 85 egg-positive ovitrap paddles were reared to adulthood for species identification and molecular analysis, yielding 682 adult mosquitoes, including 663 Aedes aegypti and 19 Aedes albopictus. A total of 631 adult female Ae. aegypti were processed into 62 pools (10 mosquitoes per pool) and one pool containing a single mosquito for molecular analysis. DENV RNA was not detected in any of the 63 mosquito pools tested (0/63, 0%; 95% CI: 0.0–2.8%) (Figures 2a–d). Wolbachia was detected in 59 of 150 trapping locations (39.9%; 95% CI: 31.8–47.4%) (Figure 2e).

Spatial distribution of Wolbachia presence

The spatial distribution of Wolbachia results is shown in Figure 3. In Figure 3a, positive and negative molecular results are displayed at the sampling-location level. Positive detections occurred in several areas, including Rawa Makmur, Lempake, and Air Hitam. Figure 3b aggregates valid results to a village-level Wolbachia index, calculated as the proportion of positive results among samples from each village. Gray areas indicate villages without sampling data. Village-level dengue incidence exhibited statistically significant positive spatial autocorrelation (Global Moran’s I = 0.116; expected I = −0.017; z = 4.03; p = 0.000055), consistent with spatial clustering. In contrast, the Wolbachia index showed no statistically significant global spatial autocorrelation (Global Moran’s I = 0.036; expected I = −0.017; z = 1.60; p = 0.109), consistent with a spatially random pattern.

For dengue, the Getis-Ord Gi analysis* (Figure 4a) identified a hot-spot cluster significant at the 99% confidence level across the north-central villages (Gunung Lingai, Sempaja Barat, Sempaja Timur, and Sempaja Selatan, and Gunung Kelua), grading to 95% and 90% confidence in the surrounding villages (including Sungai Pinang Dalam, Temindung Permai, Sidodadi, Bandara, Air Hitam, Air Putih, and Mugirejo), and a cold-spot cluster significant at the 99% confidence level at Sungai Siring, Loa Buah, and Sengkotek. The Local Moran’s I (Figure 4b) identified a north-central high–high cluster of contiguous villages, three low–low villages (Sungai Siring, Loa Buah, Sengkotek), Simpang Pasir as a high–low outlier, and low–high outlier(s) adjacent to the high–high core.

For the Wolbachia index (Figures 4c, d), only localized structure was present, consistent with the non-significant global statistic. Getis-Ord Gi* showed a north-central hot-spot grouping (Sempaja Barat, Sempaja Selatan, Air Putih, Sidodamai, Selili and adjacent villages) at the 90–99% confidence levels and a small southern cold-spot grouping (Harapan Baru and Sindang Sari). The Local Moran’s I identified a north-central high–high group (approximately eight villages, including Gunung Lingai, Mugirejo, Sungai Pinang Dalam, Bandara, Sidodadi, Gunung Kelua, Air Putih, and Sidodamai), about three low–high outliers (Sempaja Selatan, Temindung Permai, and the Sidomulyo/Bugis area), Simpang Pasir as a single high–low outlier, and Harapan Baru as a single low–low cluster.

Co-occurrence of Wolbachia and dengue clusters

Comparison of the maps showed spatial overlap in parts of central Samarinda. Sidodadi, Bandara, and Gunung Lingai were located within or adjacent to areas with elevated dengue and Wolbachia values. Mugirejo and Gunung Lingai also showed high values for both variables in the village-level maps. Simpang Pasir was identified as a high-low spatial outlier in the mapped analyses, whereas Sungai Siring and Loa Buah were located in lower-value areas for both dengue and Wolbachia. These findings describe spatial co-occurrence only and do not identify the direction, mechanism, or timing of any relationship.

This study combined citywide surveillance, entomological sampling, molecular testing, and spatial analysis in a persistently endemic setting. The findings of this study show that Samarinda City was experiencing a state of hyperendemicity in 2024. More than 70% of the administrative villages were experiencing significant dengue transmission, with the proportion of high-endemic villages being three times greater than that of endemic villages. The average infection rate of 150 per 100,000 population in some these high-burden villages far surpasses the epidemic threshold commonly utilized in national dengue surveillance (32). These observations are consistent with previous distribution patterns reported in the region (5, 33). In the context of urban Indonesia, dengue distribution typically exhibits clustered spatial patterns, heavily concentrated in densely populated neighborhoods and areas undergoing rapid land-use changes (34, 35).

The negative DENV results should be interpreted as an absence of detectable viral RNA in the tested samples, not as confirmation that vertical transmission was absent from Samarinda. This observation contradicts from multiple studies across Indonesia that have confirmed ovarian infection with DENV (21–25), as well as studies in Brazil that detected VT, albeit at low levels (36). However, systematic reviews have highlighted the significant variability of VT rates in field settings (9, 37). This difference could be explained by several factors. First, the viral loads in the collected samples may have been below the detection limits of the molecular assays employed (38). Second, intrinsic physiological barriers within the mosquito, such as the midgut infection barrier and salivary gland escape barrier, can impede viral dissemination (39). Third, mosquitoes collected prior to the completion of the extrinsic incubation period are unlikely to contain detectable viruses in their saliva (38, 40). Finally, it is generally accepted that VT is an infrequent phenomenon that contributes minimally to outbreak dynamics, particularly in hyperendemic regions where horizontal transmission (human–mosquito–human) predominates (41). Consequently, rare vertical-transmission events may have been missed, particularly if the tested sample size was limited.

The absence of VT in our study may also be attributed to the intensity of local public health interventions. In Samarinda, the notification of a Dengue Hemorrhagic Fever (DHF) case triggers immediate epidemiological investigations, followed by two cycles of thermal fogging within a 200-meter radius and rigorous mosquito breeding site elimination. Previous research in Makassar demonstrated that integrating thermal fogging with standard control measures could reduce dengue incidence by 83% (42). We assumed that the negative DENV findings are a combined outcome of the inherently low frequency of VT and the reduction of the vector population achieved through these ongoing control programs (43). Furthermore, the presence of Wolbachia in the local mosquito population may play a suppressive role. Studies have shown that w Mel-infected Ae. aegypti exhibit reduced potential for VT (12, 44). Consequently, the absence of VT in this study likely reflects not only the natural rarity of the phenomenon (45) but also the viral blocking mechanism of Wolbachia, which limits viral maintenance across mosquito generations.

Although Samarinda is not an official release site, 39.9% of the Ae. aegypti tested samples were found to be infected with Wolbachia. The origin of Wolbachia detected in Samarinda remains uncertain. Given that stable natural infections of Wolbachia in Ae. aegypti are rare in the absence of manipulation (46, 47), it is plausible that this prevalence results from a “spillover effect” from Bontang City, a Wolbachia pilot project site located approximately 100 km away (48). Anthropogenic movement and the passive transport of mosquitoes via human transportation routes could have facilitated this spread, contributing to the establishment of Wolbachia in the local vector population.

The spatial analysis revealed a complex relationship between Wolbachia presence and dengue endemicity. We observed that zones exhibiting both endemic hotspots and established Wolbachia presence continue to experience persistent transmission. This aligns with evidence that Wolbachia deployment requires near-complete population saturation to fully suppress dengue transmission (10, 41).

Our findings that zones exhibiting both endemic hotspots and established Wolbachia presence continue to experience persistent transmission align with evidence that Wolbachia alone may not fully suppress dengue without achieving near-complete saturation. Local health authorities implement focused fogging and source reduction in response each reported Dengue Hemorrhagic Fever (DHF) case. While fogging reduces adult mosquito populations, it cannot eliminate infected eggs and larvae, whereas source reduction interrupts this cycle by removing breeding habitats. Modeling studies suggest that excluding vertical transmission from control strategies may compromise intervention outcomes (49). Integrating these measures with spatial analysis of case distribution can enhance policy implementation by enabling more targeted fogging and systematic environmental management, thereby improving the sustainability of dengue control efforts.

This study found spatial heterogeneity in Wolbachia distribution across Samarinda City, with central districts displaying clusters of high prevalence; peripheral zones showing low prevalence despite bordering high-prevalence areas, implying possible barriers to Wolbachia propagation or localized gaps in deployment; and certain areas with high Wolbachia presence nested within low-prevalence surroundings. These spatial anomalies may stem from several factors, including the nonlinear dynamics of mosquito populations influenced by density-dependent competition, imperfect maternal transmission of Wolbachia, and physical or infrastructural impediments that restrict mosquito movement and impede uniform spread. Cluster-randomized trials in Yogyakarta and field studies in Malaysia demonstrated significant reductions in incidence yet also emphasized that effectiveness depends on high Wolbachia prevalence and stable establishment in mosquito population (13, 48). Although Wolbachia -based strategies have shown promise in reducing vector competence, the extent to which their establishment interacts with environmental variability and localized transmission hotspots remains insufficiently understood.

The spatial alignment of infection rates with the endemic and Wolbachia maps reveals a strong correlation: high infection rates persist where both endemicity is high and Wolbachia intervention is either insufficient or still maturing. In contrast, low infection rates are found in either natural cold spot regions or where Wolbachia has shown early effectiveness. Spatially explicit models incorporating realistic habitat heterogeneity and density-dependent competition show that Wolbachia spread slowly in uneven environments, especially where juvenile competition varies spatially, and benefits from immigration from neighboring areas. Such persistence could be attributable to incomplete mosquito population saturation, vector resistance, delayed intervention effects, suboptimal coverage, or external factors including adverse environmental conditions and human mobility. Conversely, areas characterized as cold spots for both endemicity and Wolbachia likely benefit from naturally inhospitable environments or low population density, which serve as buffers against dengue transmission. The findings of the study also showed that as many as 62.4% of the 150 catch points were found by Wolbachia mosquitoes. Wolbachia efficiently reduces the proportion of mosquitoes that reach the potential for infection and transmission, especially in low and medium endemic (50, 51).

Strengths and limitations

A major strength of this study is the integration of a complete citywide dengue-surveillance dataset with geocoded household sampling, mosquito identification, molecular testing, and multiple spatial-statistical approaches. The citywide epidemiological analysis included all laboratory-confirmed dengue cases recorded in the 2024 surveillance database, while the entomological survey included all eligible index households identified during field surveillance. This design provides a comprehensive framework for linking human dengue cases with entomological findings at an operationally relevant geographic scale.

Several limitations should be considered. The entomological survey covered a subset of case households during a four-month period, and its representativeness depends on the household-selection procedure. The probability of detecting rare DENV-positive mosquito or egg samples was limited by the number and timing of specimens, pooling strategy, and assay detection limit. The study did not measure Wolbachia strain, bacterial density, mosquito-population coverage, duration of establishment, vector abundance over time, environmental conditions, human mobility, or intervention intensity. The village-level spatial analysis is ecological and cannot establish individual-level or causal relationships. These limitations constrain inference regarding vertical transmission, the origin of Wolbachia, and its effect on dengue transmission.

Samarinda experienced a high and spatially heterogeneous dengue burden in 2024, with 67.8% of villages classified as high endemic. DENV was not detected in the adult female mosquito or egg samples tested; this finding provides no evidence of vertical transmission in the tested samples but does not exclude rare transmission in the wider mosquito population.

Wolbachia was detected at a substantial proportion of evaluated locations and overlapped spatially with some dengue hotspots. Because the study did not measure strain identity, density, population coverage, duration, or a counterfactual comparison, these maps cannot establish the origin or effectiveness of Wolbachia. Repeated, standardized molecular and entomological surveillance linked to spatial epidemiology is needed to clarify these relationships and to guide targeted, integrated vector control.