1. Introduction

The dengue virus (DENV), a flavivirus with four distinct serotypes, poses a significant and growing threat to global public health, with an estimated 390 million infections occurring annually [1]. The clinical spectrum ranges from asymptomatic infection to life-threatening dengue hemorrhagic fever and dengue shock syndrome. Historically, the primary symptoms have been fever, bleeding, and fluid leakage from blood vessels.

However, emerging evidence challenges this traditional understanding. Recent outbreaks demonstrate that dengue’s clinical spectrum extends well beyond classical hepatic and hematological manifestations. Systematic reviews of clinical and radiological data reveal that pulmonary involvement, though historically underappreciated, represents a significant component of dengue pathophysiology [2].

Pulmonary involvement in dengue has been increasingly documented across diverse geographic regions and serotypes. A systematic analysis of recent literature reveals consistent patterns: Rodrigues and colleagues (2014) documented pleural effusion in 55.2% and ground-glass opacities in 27.6% of 2,020 Brazilian patients with predominantly DENV-1 and DENV-2 infections [3]. Similarly, Almeida and colleagues’ (2017) systematic review spanning 56 years established a hierarchy of pulmonary manifestations: pleural effusion (17%–100%), noncardiogenic pulmonary edema (2%–29%), and diffuse alveolar hemorrhage (<1%) [2]. More recent hospital-based cohorts provide clinical context: Gupta and colleagues (2020) reported 56% pulmonary complications in an Indian ICU setting with 22% developing acute respiratory distress syndrome [4], while Keshav and colleagues (2024) documented lower prevalence (10.2%) but higher mortality (23%) among those with pulmonary involvement [5]. Notably, these studies predominantly examined DENV-1, DENV-2, and DENV-3, highlighting a critical gap in DENV-4-specific respiratory manifestations that our study addresses.

These studies reveal a critical knowledge gap: although pulmonary manifestations have been documented in DENV-1, DENV-2, and DENV-3 infections, the literature lacks systematic characterization of DENV-4-associated respiratory complications, especially in primary outbreaks in populations with no prior exposure. This gap is particularly important given the different immunopathological profiles and clinical severity patterns observed among the serotypes. These findings suggest that DENV may have unique pulmonary tropism characteristics that require further investigation.

In 2023, Lincang City in Yunnan Province, China, experienced a notable dengue fever outbreak caused by Dengue Virus Serotype 4 (DENV-4) (Fig 1A). Yunnan has historically been recognized as a dengue hotspot in mainland China, partly due to its favorable climatic conditions that support the breeding of Aedes mosquitoes [1]. While previous studies have primarily emphasized the hemorrhagic manifestations and hepatic or hematological complications of dengue, the 2023 outbreak emphasized the importance of pulmonary symptoms in disease progression. This viewpoint synthesizes findings from 147 confirmed cases in Lincang and emphasizes why pulmonary symptoms warrant more clinical and research attention. By drawing upon newly published studies and contextualizing the outbreak within the evolving clinical profile of dengue, we aim to provide insights that are relevant to clinicians, public health professionals, and researchers of neglected tropical diseases.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Epidemiological, temporal, clinical, and indicator symptoms for severe dengue of the 2023 dengue fever outbreak in Lincang, China.

A, Geographic location of Lincang. Base map: U.S. Geological Survey (public domain), usgs.gov. Map compilation and annotations by the authors. B, Temporal trend of the dengue fever outbreak. C, Sex-specific differences in clinical symptoms. D, Indicator symptoms for severe dengue fever. E, Visualization of the symptom network. This figure presents a graphical representation of the symptom network derived from dengue fever cases. Only statistically significant associations between symptoms (p < 0.05) are included, with nodes representing symptoms and edges indicating their correlations. The weight of the edges reflects the strength of the associations. Refer to Table 1 in S1 Table for the symptom codes. F, Centrality measures of symptoms in the network.

https://doi.org/10.1371/journal.pntd.0013679.g001

2. Ethics statement

Formal written consent was obtained from all participants and their parents/guardians prior to their involvement in the study. The consent process ensured that both the participants and their parents/guardians were fully informed about the purpose, procedures, and potential risks of the study. The ethical approval was provided by the Yunnan Center for Disease Control and Prevention (No.QR-KJB-2021-003).

3. Laboratory confirmation and serotype assignment

The virological diagnosis was conducted in accordance with the National Laboratory Testing Scheme for Dengue, as stipulated by the China CDC (2023) [6,7]. Acute-phase sera (≤5 days post-onset) underwent NS1 antigen capture immunochromatography (Wantai BioPharm) and multiplex real-time RT-PCR targeting the C-prM junction using serotype-specific probes (Shengxiang Biotechnology, CFDA No. SXDD-04). All RT-PCR-positive samples (n = 147) yielded DENV-4 amplification with Ct values ranging from 18.3 to 36.7 (median 28.4).

In order to satisfy the national demand for molecular epidemiological surveillance during outbreaks, E-gene sequencing (1,485 bp) was performed on all specimens. Phylogenetic analysis employing the MEGA 11.0 software program confirmed that all sequences fell within the DENV-4 genotype I cluster.

For convalescent sera with a duration of infection greater than five days or for cases in which the results of the reverse transcription-polymerase chain reaction (RT-PCR) were inconclusive, a paired immunoglobulin M (IgM)/IgG enzyme-linked immunosorbent assay (ELISA) was performed using the Panbio kit. The presence of seroconversion or a fourfold increase in titer was indicative of an acute infection. The quality assurance program encompassed the validation of the specimens by a provincial reference laboratory and the utilization of national standard positive/negative control panels. The analysis was conducted exclusively on cases that satisfied the criteria for dual-assay confirmation (S1 Text).

4. Epidemiological context

Between August and October 2023, 147 cases of dengue fever (all confirmed to be DENV-4 infections) were hospitalized at the People’s Hospital of Lincang, a city located near the Myanmar border (Fig 1B). Epidemiological investigations indicated that 51.7% (76/147) of the patients were male, with a median age of 43 years (interquartile range [IQR]: 33–58). Approximately 76.2% of the cases originated from Gengma County, a region frequently noted for its dense Aedes mosquito population [8]. These demographic parameters are consistent with prior dengue incidence trends in Yunnan, although the 2023 outbreak demonstrated unique clinical features that differentiate it from previous episodes in the same area [9].

Key contributing factors to the increasing dengue incidence in these regions include rising temperatures, climatic changes, increased mobility, and cross-border movement [10]. Indeed, the spillover of dengue outbreaks across neighboring countries has been observed, facilitated by porous borders and fluctuating mosquito vector burdens. This underscores the complexity of controlling dengue, as any local outbreak could potentially assume a regional dimension. Furthermore, the COVID-19 pandemic has likely exacerbated diagnostic challenges, as limited healthcare resources and overlapping symptoms may have contributed to delayed dengue detection in some patients [11].

5. Clinical presentation and methodological overview

All 147 hospitalized cases underwent standardized clinical and virological evaluation. Comorbidities were classified using to ICD-10 codes, and disease severity was classified per established guidelines (Table A in S1 Table). Notably, the median hospitalization duration was seven days (IQR: 6–9), and the median maximum body temperature reached 39.0°C (IQR: 38.1–39.3°C) (Table B in S1 Table). These values align with typical dengue presentations seen globally, although the range of “febrile phases” extends considerably among patients [12].

Pulmonary assessment was conducted in accordance with the 2023 National Dengue Prevention and Control Protocol, employing a tiered approach. At admission and subsequently on a daily basis, all patients underwent a standardized respiratory screening procedure. This procedure included a review of symptoms such as cough, dyspnea, chest pain, and sputum. In addition, a physical examination with auscultation was performed, and the respiratory rate and room-air SpO₂ were recorded. When receiving oxygen, the patient’s position and the FiO₂ were documented. The diagnostic approach was meticulous and judicious, entailing a sequential process of evaluation. Initially, chest X-rays were obtained to ascertain the presence of respiratory symptoms, abnormal lung findings, or hypoxemia. Subsequently, when feasible, handheld point-of-care lung ultrasonography was employed to assess interstitial changes and pleural effusions. Finally, chest CTs were reserved for cases that met predefined criteria, such as severe, non-resolving, or atypical cases, where the results were anticipated to lead to a modification in management.

Patients were monitored daily for symptom evolution. Resolution was defined as the first 24-hour period free of chest distress, dyspnea, cough and expectoration without the need for supplemental oxygen or bronchodilators. Time-to-resolution (TTR) was calculated from hospital admission to symptom clearance.

We employed EBICglasso network analysis to identify symptom relationships. This method uses regularized partial correlations, which differ from simple correlations by controlling for all other symptoms simultaneously [13]. For example, the correlation between chest distress and dyspnea (edge weight = 0.23) represents their direct association after removing the influence of all other symptoms. The LASSO penalty eliminates spurious connections, ensuring only robust associations remain. Consequently, the emergence of the pulmonary cluster with edge weights ranging from 0.19 to 0.23, although seemingly modest, signifies the presence of robust conditional dependencies that have endured rigorous statistical filtration. From a clinical perspective, this grouping of symptoms indicates that they represent a cohesive syndrome rather than distinct manifestations. This finding is significant because it has substantial implications for screening methodologies and management strategies (see Fig 1E and 1F).

6. Sex-specific patterns

Sex-specific differences led to interesting observations. Females more frequently reported weakness (93.0%), leukopenia (63.4%), headaches (53.5%), and hypokalemia (40.8%), while males showed higher tendencies for hyperuricemia (22.4%) and splenomegaly (11.8%). The incidence of splenomegaly in dengue, though not the most classical finding, can occur and should not be disregarded when assessing patient outcomes (Fig 1C, Table C in S1 Table).

An accumulating body of evidence indicates that sex differences in dengue reflect intertwined biological and socio-behavioral processes rather than a single causal mechanism. On the biological axis, estrogens modulate endothelial tone and barrier integrity through nitric oxide-dependent and receptor-mediated signaling, plausibly influencing the propensity to plasma leakage that underlies severe disease, while broader sexual dimorphism in immune responses can shape viremia kinetics, inflammatory profiles, and downstream clinical manifestations [14].

Experimental and translational studies demonstrate that estrogens enhance endothelial resilience and reduce paracellular permeability, with potential implications for pulmonary and serosal fluid shifts during critical-phase dengue, whereas females generally mount more robust innate and humoral responses, which may alter both susceptibility and phenotype expression across the disease course.

It is hypothesized that socio-behavioral determinants likely operate in parallel with biological mechanisms, contributing to sex-differential dengue outcomes. It has been demonstrated that men may incur greater vector exposure through occupational and outdoor activities. In contrast, women in some settings face delayed access to formal care or preferential reliance on traditional providers. These dynamics have the potential to generate surveillance artifacts and presentation delays, which can lead to the inflation of apparent clinical disparities that are not necessarily indicative of biological differences.

A pre-specified framework will differentiate these pathways by: The integration of sex- and stage-specific factors (menstrual/menopausal status, pregnancy, exogenous hormone use) with immune and endothelial biomarkers is a crucial aspect of the study. The quantification of exposure and access variables, including but not limited to occupation, outdoor time, protective practices, household vector control, time-to-presentation, and decision-making norms, is essential for a comprehensive understanding of the subject matter. The application of sex-stratified models, incorporating interaction terms and causal mediation, aims to partition total effects into biological and socio-behavioral components. This approach is supplemented by sensitivity analyses for unmeasured confounding and differential misclassification in symptom reporting or care-seeking.

From a clinical perspective, a dual-action strategy is recommended: first, implement sex-aware respiratory risk appraisal and conservative fluid stewardship for patients with lower-respiratory symptoms or hypoxemia (reflecting potential biological risk); second, concurrently strengthen outreach and triage access to reduce care-seeking delays and exposure-driven risk. To address this gap, routine equity monitoring should be implemented to detect and remediate sex-differential gaps in processes and outcomes.

7. Pulmonary symptoms: An emerging concern

Pulmonary symptoms were present in 46 patients (31.3%). The most frequent complaint was cough (38/147, 25.9%), followed by dyspnea (29/147, 19.7%), and chest distress (22/147, 15.0%).

The respiratory symptoms resolved rapidly under the administration of supportive care. The median TTR was found to be 4 days (IQR 3–6) overall and 6 days (IQR 5–8) in severe cases. A total of 46 symptomatic patients demonstrated complete resolution by day 14 of follow-up, with no evidence of residual pulmonary sequelae.

The prevalence of lower respiratory symptoms in the Lincang outbreak appears to be higher than that observed in general ward-based dengue cohorts, but lower than that seen in severe/critical care strata. In Brazil, Rodrigues and colleagues (2014) evaluated a highly selected subset of 29 dengue patients who underwent chest CT because of respiratory complaints and reported cough in 27.6% and dyspnea in 37.9% within that CT subset; only 1.43% of the broader 2,020 case cohort underwent CT, so these figures are not population prevalences [2]. In India, Gupta and colleagues (2020) found that among hospitalized dengue patients with pulmonary complications, dyspnea and cough occurred in 92.8% and 78.6%, respectively, compared with 9.0% and 54.5% in those without pulmonary complications [4]. In a larger Indian inpatient cohort, Keshav and colleagues (2024) observed overall cough and dyspnea rates of 16.1% and 5.5% across 255 patients, rising to 26.9% and 30.8% among those with lung involvement [5]. In this context, the DENV-4 figures in Lincang (25.9% for cough and 28.6% for LRI) exceed the routine ward level averages and are comparable to selected imaging-triggered series. However, they remain well below the frequencies observed in severe disease with pulmonary complications [2,4,5].

The network analysis revealed that pulmonary symptoms form a distinct cluster with edge weights (regularized partial correlations) of 0.19–0.23. These values, which represent conditional associations subsequent to controlling for all other symptoms, indicate robust direct relationships within this syndromic complex. Despite the presence of moderate individual edge weights, the emergence of the cluster indicates a unified pathophysiological process that possesses substantial prognostic value. Keshav and colleagues reported an odds ratio of 6.7 (95% CI 3.2–13.9) for “positive lung imaging to death” [5]. Collectively, the data delineate a step-up pattern from symptoms to imaging positivity to fatality, thereby emphasizing the feasibility of early symptom-based triage.

Our serotype-specific data and symptom-network analysis complement radiological and autopsy literature, indicating that “pulmonary linkage” is an evolving component of the dengue clinical spectrum rather than a mere complication. From a pathophysiological standpoint, pulmonary involvement in dengue likely arises from a synergistic interplay between the virus’s direct effect on capillary and endothelial function, immune-mediated inflammatory cascades, and possible secondary bacterial or viral infections [15]. When alveolar-capillary integrity is compromised, fluid leakage into pulmonary interstitial can precipitate acute respiratory distress. This phenomenon underscores the importance of early intervention, including fluid management, oxygen supplementation, and strict monitoring of respiratory parameters, particularly in settings where dengue presentations have diversified beyond the classical febrile-hemorrhagic spectrum.

8. Predictors of severe disease

Univariate and multivariable logistic regression analyses pinpointed thrombocytopenia and hypoplasia (diminished liver function) as significantly associated with severe dengue fever (Fig 1D, Table D in S1 Table). Thrombocytopenia (OR = 3.06, 95% CI: 1.05–9.03) emerged as a well-known and robust predictor, aligning with longstanding research on platelet count as a core marker of disease severity in dengue [16]. Hyperpathia (OR = 3.82, 95% CI: 1.00–13.61) likewise flagged disease severity, consistent with previous outbreaks where hepatic involvement was a key indicator of poor outcomes [17]. Notably, lower respiratory infection coexisting with dengue had an OR of 2.91 (95% CI: 1.05–8.07) in univariate analysis, further highlighting the significance of respiratory system stress in severe clinical phenotypes.

Potential interactions between systemic and pulmonary manifestations remain incompletely understood. For example, the synergy between hepatic dysfunction and inflammatory cascades could predispose alveolar and bronchial tissues to exudative processes. Although the current study’s sample size was modest (n = 147), the strength of associations between pulmonary symptoms and severity underscores the need for further large-scale research. It also signals that comprehensive assessment of respiratory parameters (e.g., imaging studies, arterial blood gases) might become more standard in resource-limited settings wherein dengue remains endemic.

9. Implications for clinical management and research

Lower-respiratory manifestations of dengue, including chest discomfort, dyspnea, cough, and sputum production, should be explicitly integrated into routine assessment and triage within existing dengue pathways. These features serve as clinically accessible indicators of systemic inflammation and plasma leakage, which can be detected using scalable, low-cost tools such as standardized symptom inquiry and pulse oximetry. Consequently, they are well-suited for implementation in resource-limited settings.

10. Conclusion

This study fundamentally challenges our understanding of DENV-4 clinical presentation. The high prevalence of pulmonary manifestations (43.5%), their central role in the symptom network, and strong association with disease severity demand immediate integration into clinical protocols. While our findings require validation in larger, multi-center studies, the evidence compels a paradigm shift: respiratory assessment must become a cornerstone of dengue management, not an afterthought. As climate change expands dengue’s geographic reach, recognizing and managing these pulmonary manifestations will be critical for reducing morbidity and mortality across endemic regions.

However, these conclusions are tempered by important limitations. The analysis derives from a modest sample within a single outbreak, and imaging and microbiologic testing were symptom- and resource-triggered rather than universal. This approach potentially enriches the sample for severe phenotypes and introduces selection bias and misclassification. Consequently, the findings warrant cautious interpretation and prospective validation. The following areas have been identified as priorities: multicenter studies to confirm associations, elucidate the pathophysiological links between systemic inflammation, plasma leakage, and respiratory phenotypes, and refine therapeutic approaches that balance conservative fluid strategies with timely oxygen support and judicious antimicrobial use. Proactive research, in conjunction with integrated, respiratory-focused monitoring, will be essential to mitigate complications and reduce dengue-related mortality as clinical presentations continue to evolve.

Supporting information