Platelet dynamics and thrombocytopenia in dengue fever: A prospective cohort study from Shenzhen, China

Abstract

Introduction

Dengue fever, the most prevalent arthropod-borne viral disease, causes ∼400 million infections annually. Although thrombocytopenia is commonly associated with dengue, how it evolves in relation to viral load and immune responses remains poorly understood. This study aimed to elucidate platelet-virus-immune interactions in acute dengue by systematically tracking of viral load, platelet parameters, and leukocyte dynamics.

Methods

A prospective cohort study was conducted at Third People's Hospital in 2024, involving 135 confirmed dengue cases, supported by retrospective data from 2014 to 2023. Platelet counts, hematocrit (HCT), and cellular immunity markers (lymphocyte/neutrophil percentages) were longitudinally tracked. Viral load was quantified via NS5 gene Ct values. Statistical analyses involved LOESS regression and Pearson/Spearman correlations.

Results

Platelet counts exhibited a biphasic decline, reaching nadir levels (mean: 97.65 × 10⁹/L) at 6 days post-onset, with recovery by day 9. Thrombocytopenia severity was stratified as intermediate-low (50–99 × 10⁹/L; 50 %, 64/128) and very low (<50 × 10⁹/L; 14.8 %, 19/128). Platelet decline correlated with elevated lymphocyte percentages (40 % vs. 17.8 % pre-decline; p < 0.001) and suppressed neutrophils (46.6 % vs. 68.3 %; p < 0.001). Critically, platelet counts inversely correlated with viral load (Ct values: R = 0.25, p = 0.028), HCT (R = −0.25), and platelet activation markers (MPV: R = −0.55; P-LCR: R = −0.57), while positively associating with platelet hematocrit (PCT: R = 0.97). No cases progressed to severe dengue despite extreme thrombocytopenia.

Conclusions

This study identifies distinct dengue thrombocytopenia kinetics driven by viral load. Predominant moderate thrombocytopenia (50–99 × 10⁹/L) challenges conventional risk stratification, advocating integrated monitoring of platelet indices and viral replication. These data advance both risk prediction and mechanistic knowledge of platelet-virus interactions.

Highlights

• •Biphasic platelet kinetics in dengue, with 6-day nadir and day-9 recovery pattern.
• •Moderate thrombocytopenia (50-99 × 10⁹/L) predominates, challenging current severity criteria.
• •NS5 Ct values link viral load to thrombocytopenia and lymphocyte surge.
• •Platelet activation markers (MPV/P-LCR) show stronger correlation than hematocrit.

1. Introduction

Dengue (DENV) represents the most prevalent arthropod-borne viral infection worldwide, with an estimated 400 million cases annually, of which 96 million are symptomatic, resulting in over 40,000 deaths [[1], [2], [3]]. Despite considerable advancements in our comprehension of the virus, encompassing its clinical characteristics, immune response, and disease progression, as well as the implementation of various interventions such as effective clinical management, partially effective vaccines, vaccines in development, and innovative mosquito control strategies, the global burden of dengue has continued to escalate over the past decade [3,4]. Projections indicate that the global population at risk is expected to rise from 53 % in 2015 to 63 % by 2080, with high environmental suitability for dengue persisting in tropical and subtropical regions worldwide [5]. This trend suggests a substantial and ongoing burden on healthcare resources, which may be further exacerbated during periodic outbreaks [1]. Notably, Guangdong Province bears the highest dengue burden in China, accounting for over 80 % of national cases, with Shenzhen serving as a major local population center and international gateway. For example, surveillance data from Shenzhen in 2014 documented 454 dengue cases, of which 76.2 % were locally acquired and 23.8 % were imported (primarily from Southeast Asia and neighboring cities). The cases exhibited strong seasonal clustering, with 97.1 % occurring between September and November, with adults aged 20–50 years representing 76.7 % of cases [6].

Although dengue is primarily a self-limiting acute febrile illness characterized by non-specific symptoms, approximately 60–80 % of individuals infected with the dengue virus (DENV) remain asymptomatic or experience subclinical infections [3], a substantial proportion of patients require hospitalization due to high fever or other severe symptoms, particularly when warning signs are present. Thrombocytopenia is a prevalent complication among hospitalized dengue patients, often accompanied by an increase in hematocrit (HCT) levels, which typically occurs concurrently with a rapid decline in platelet count. [3,[7], [8], [9], [10]]. According to the World Health Organization's 2009 guidelines, a rapid decline in platelet count concurrent with a rise in HCT is considered a laboratory warning sign of clinical significance, comparable to mucosal bleeding [9]. The 2024 edition of China's diagnosis and treatment guidelines introduced a platelet threshold of <50 × 10⁹/L, further refining this framework (https://www.gov.cn/zhengce/zhengceku/202407/P020240729756402615362.pdf). Clinicians are understandably concerned about risk of severe disease with low platelet and poor platelet recovery, and it is clear that not all patients are candidates for prophylactic platelet transfusion after the occurrence of thrombocytopenia also has certain indications [[11], [12], [13]]. However, although thrombocytopenia is a hallmark of dengue infection with significant prognostic implications, its underlying mechanisms remain poorly understood. Critical knowledge gaps include: (1) the lack of characterized quantitative relationships between dengue viral load (NS5 Ct values) and thrombocytopenia kinetics, and (2) the undefined combined predictive utility of platelet indices (MPV/P-LCR) with immune markers for thrombocytopenia progression. Furthermore, data on the dynamics of platelet responses and the clinical consequences of severe thrombocytopenia in dengue are limited.

To address these gaps, this prospective longitudinal study longitudinally tracked NS5 Ct values, platelet indices (MPV/P-LCR), and immune markers to define their dynamic interrelationships and clinical thresholds for severe thrombocytopenia.

2. Methods

2.1. Patient information and sample collection

The cohort for this study comprised patients hospitalized at Shenzhen Third People's Hospital from 2014 to 2024. Data were harmonized retrospectively (2014–2023) and prospectively (2024) using consistent clinical and laboratory protocols. Comprehensive information, including clinical symptoms and laboratory examination results, was gathered at the earliest possible time following patient admission. Blood samples were obtained at intervals of 1–3 days throughout the hospitalization period. All samples were collected in EDTA tubes and processed within 2 h of venipuncture to reduce pre-analytical effects on platelet indices. To ensure data accuracy, two independent researchers conducted cross-validation of data entries.

2.2. Laboratory confirmation of dengue infection

The confirmation of dengue virus infection is achieved through the detection of the dengue virus NS1 antigen in the blood during the acute febrile phase, utilizing colloidal gold or enzyme-linked immunosorbent assay (ELISA) methods, as recommended by guidelines for the diagnosis and treatment of dengue fever. Exclusion criteria encompassed co-infection with other flaviviruses, such as Zika or Chikungunya, pre-existing hematologic disorders, such as idiopathic thrombocytopenic purpura, and recent administration of immunosuppressive agents. Dengue virus-specific IgM and IgG antibodies were identified in acute-phase serum samples via colloidal gold-based immunochromatographic rapid diagnostic tests (Guangzhou Wondfo Biotech), following the manufacturer's protocol. The determination of virus genotypes and viral cycle threshold (Ct) values for the NS5 gene was conducted using probe primers supplied by the Centers for Disease Control (CDC) of Shenzhen, China, in conjunction with sequencing and epidemiological data. Platelet grading and stratification were based on previously established criteria [14]. Abnormal platelet counts (PLT <50 or >400 × 10⁹/L) were verified through manual microscopic examination, and viral load was assessed using the mean of three repeated measurements. All patients were monitored until discharge. The day following the onset of the first symptoms was designated as day 0 post illness onset for DENV infections. The clinical features of patients were monitored every 1–3 days during hospitalization, and peripheral blood and serum samples were collected for hematological and biochemical analyses.

2.3. Clinical symptom and hemodynamic parameter analysis

Clinical symptoms were systematically documented throughout hospitalization. Comprehensive Hemodynamic evaluations were performed, including white blood cell counts, neutrophil proportions, lymphocyte proportions, monocyte proportions, HCT, and platelet counts in dengue virus-infected patients. These parameters were assessed in relation to both symptom onset and laboratory confirmation. Platelet count dynamics were longitudinally monitored across the entire infection course. Additionally, the influences of potential modifying factors on platelet dynamics were examined, including age, gender, and underlying health conditions (chronic liver diseases, diabetes, hypertension, and pregnancy status).

2.4. Statistical analysis

For descriptive analyses, data were presented as medians (interquartile range, IQR) for continuous parameters and as frequencies (percentages) for categorical variables. T-test and Wilcoxon rank-sum test were used to compare the continuous variables for normal and non-normal data, respectively. The chi-square test and Fisher exact test were used to compare count data. The 95 % CI for clinical data were calculated as odds ratios using Fisher's exact test. Data sets were considered significantly different if the P value was less than 0.05. Correlations of platelet counts and platelet related markers as well as Ct values were calculated using linear fitting regression as well as Pearson correlation coefficients. We calculated absolute risks using the formula:

AR = Events/Total cases × 100 %

What's more, the Hemodynamic parameters were calculated using LOESS (locally estimated scatter smoothing) curve fitting through polynomial regression. Statistical analyses were conducted using R version 4.1.0 and GraphPad Prism version 8.0.1.

2.5. Ethical considerations

This study was approved by the Ethics Committee of Shenzhen Third People's Hospital (2021-002). All participants in the prospective cohort provided written informed consent, while retrospective data analysis used anonymized clinical records in accordance with institutional ethical guidelines and national regulations.

3. Results

3.1. Kinetics and specificity of clinical and hemoimmunodynamic characteristics

Patients with laboratory-confirmed DENV-2 infection who were admitted to our hospital between April 5th, 2024, and 30th, 2024, were included in this study, totaling 135 cases. During the first 7 days of illness, 87 out of 135 patients (64.44 %) underwent dengue IgM/IgG serological testing, yielding 2 IgG-positive cases (2.30 %) and 66 IgG-negative cases (97.70 %). A retrospective review found no clinical evidence of prior dengue exposure in the IgG-positive patients. During the initial outbreak phase (April 2024), all confirmed cases were travel-associated infections (n = 5), comprising two patients from Indonesia (n = 2), two from Malaysia (n = 2), and one from Bali (n = 1). However, starting in July, the proportion of locally acquired infections (including Shenzhen and neighboring cities) gradually increased and eventually became the dominant transmission pattern (Fig. 1A). The proportion of male patients (68.89 %, 93/135) was higher than that of female patients. The median age of all participants was 37 years (IQR: 28–53), with both genders predominantly distributed in the 15–34 and 35–64 age groups (Fig. 1B). The top 5 frequently reported symptoms for the DENV infections were fever (98.52 %, 133/135), Dizziness or/and headache (77.04 %, 104/135), Fatigue (68.89 %, 93/135), Muscle or/and joint pain (64.44, 87/135), Rash (19.26 %, 26/135) (Fig. 1C). Notably, fever was common across all age groups; however, the incidence of Muscle or/and joint pain and Dizziness or/and headache (28.71 %, 1/7) were much lower in patients younger than 14 years than in the other three age groups, and no rash was observed in this age group. In contrast, elderly patients (≥65 yr) exhibited a higher rate of Fatigue (84.62 %, 11/13). The prevalence of Muscle or/and joint pain and Dizziness or/and headache was higher among young (15–34 yr) and middle-aged (35–64 yr) patients, with the latter group showing the highest rates (86.36 %, 57/66 and 77.27 %, 51/66, respectively) (Fig. 1C). Then we analyzed the hemocyte dynamics, including lymphocyte percentage (Fig. 1D), blood cell counts (Fig. 1E), neutrophil percentage (Fig. 1F), monocyte percentage (Fig. 1G), HCT (Fig. 1H) and platelet count (Fig. 1I) during the post-illness onset period. The results showed that the lymphocyte percentage increased during the early and middle stages of DENV infection and subsequently decreased, remaining generally within the normal range throughout disease progression. In contrast, the neutrophil percentage exhibited an opposite trend, decreasing during the early and middle stages and increasing during the late stage. Blood cell counts and monocyte percentage did not show significant changes during the illness onset period. Notably, the HCT increased slightly during the middle of the disease, although the change was minimal and remained within the normal range. However, platelet counts declined sharply starting from the third day after illness onset, reaching the lowest level on the sixth day, followed by a gradual return to the normal range by the ninth day. None of the patients in this cohort progressed to severe disease.

Fig. 1.

Clinical and hemoimmunodynamic characteristics of dengue infection. (A) Proportion of patients admitted from April to October. (B) Number of male and female patients by age group. (C) Patient symptoms categorized by age group. (D–I) Dynamics of lymphocyte percentage, white blood cell counts, percentage, monocyte percentage, hematocrit value, and platelet counts of dengue infection throughout infection course in the host. Individual data points are shown as circles; range of normal values are shown as lines, with 95 % CIs shown as shading. Abbreviations: LYMPH, lymphocyte; WBC, white blood cell; NEUT, neutrophil; MONO, monocyte; HCT, hematocrit; PLT, platelet.

3.2. Hemocyte values in specific stage of platelet declining

According to the trends in platelet changes and the occurrence of thrombocytopenia, we categorized the patient's disease course into three phases: the pre-decline phase (0–2 days), the decline phase (3–9 days), and the recovery phase (10–16 days). The mean platelet count was lowest during the decline phase (mean: 97.65 × 10⁹/L, 95 % CI: 91.81-103.5 × 10⁹/L), followed by the pre-decline phase (mean: 151.2 × 10⁹/L, 95 % CI: 138.5-163.8 × 10⁹/L) and the recovery phase (mean: 219.2 × 10⁹/L, 95 % CI: 165-277.9 × 10⁹/L) (Fig. 2A). HCT levels were significantly higher during the decline phase (42.42, 95 % CI: 41.88–42.95) compared to both the pre-decline phase (40.33, 95 % CI: 38.89–41.77) and the recovery phase (37.94, 95 % CI: 35.87–40.02) (Fig. 2B). We further analyzed the variations in white blood cell counts (Fig. 2C), lymphocyte percentage (Fig. 2D), neutrophil percentage (Fig. 2E), monocyte percentage (Fig. 2F) across the three phases. The results indicated that white blood cell counts reached a lower peak during the decline phase (mean: 3.919 × 10⁹/L, 95 % CI: 3.624-4.215 × 10⁹/L) compared to the recovery phase (mean: 5.481 × 10⁹/L, 95 % CI: 4.358-5.605 × 10⁹/L) (p < 0.001), with no significant difference observed between the decline and pre-decline phases (mean: 4.074 × 10⁹/L, 95 % CI: 3.665-4.483 × 10⁹/L). The lymphocyte percentage during the decline phase (40 %, 95 % CI: 38.08–41.91 %) was significantly higher than that in the pre-decline phase (17.80 %, 95 % CI: 14.70–20.89 %) (p < 0.001). Conversely, the neutrophil percentage during the decline phase (46.59 %, 95 % CI: 38.08–41.91 %) was significantly lower than in the pre-decline phase (68.34 %, 95 % CI: 64.43–72.34 %) (p < 0.001). Additionally, the monocyte percentage during the decline phase (11.52 %,95 % CI: 10.94–12.09 %) was slightly lower than that in the pre-decline phase (13.35 %, 95 % CI: 11.63–15.07 %) (p = 0.028).

Fig. 2.

Immune cell changes during various periods of platelet decline. Platelet counts (A), hematocrit value (B), white blood cell counts (C), lymphocyte percentage (D), neutrophil percentage (E), and monocyte percentage (F) comparison among the platelet pre-decline phase, decline phase, and recovery phase. Abbreviations: LYMPH, lymphocyte; WBC, white blood cell; NEUT, neutrophil; MONO, monocyte; HCT, hematocrit; PLT, platelet.

3.3. Effect of platelet count on cellular immunity

Next, we incorporated platelet and HCT data from 2014 to 2023 for comparison with 2024. Interestingly, the results revealed that HCT in 2024 was the highest in the past five years (Fig. 3A), whereas the platelet count was nearly the lowest in the past decade, except for 2018 (Fig. 3B). Furthermore, we categorized patients into into four groups based on their platelet counts during the critical phase: normal group (platelet count ≥125 × 10⁹/L), low group (100-124 × 10⁹/L), intermediate low group (50-99 × 10⁹/L), and very low group (≤49 × 10⁹/L). The results indicated that only 17.2 % (22/128) of patients maintained normal platelet values during the critical phase, with the intermediate low group comprising the largest proportion at 50 % (64/128), followed by the low group at 18.0 % (23/128). Notably, in this dengue fever outbreak, the very low group accounted for a considerable proportion of 14.8 % (19/128) (Fig. 3C). Additionally, compared to the normal group, patients in the low, intermediate low, and very low platelet groups exhibited significantly higher lymphocyte percentages (Fig. 3D), and significantly lower neutrophil percentages (Fig. 3E), although no significant difference were observed among these three groups. Regarding monocyte percentage, no significant difference was found between the normal group and the low groups, while the intermediate low group showed lower levels than both the normal group and the low group (Fig. 3F). To further assess whether low counts were associated with potential patient-related factors, we analyzed platelet dynamics across age groups, genders, and presence or absence of underlying diseases, and compared the minimum platelet counts during the decline phase. The results demonstrated no statistically significant differences in platelet dynamics or minimum platelet counts among these subgroups (Fig. S1).

Fig. 3.

Comparison of thrombocytopenia over the years and percentage of immune cells in each platelet count group. (A–B) Comparison of hematocrit value and platelet counts during the 2014 yr–2024 yr. (C) The proportion of the population under each group of platelet counts. (D–F) Comparison of lymphocyte percentage, neutrophil percentage, and monocyte percentage in each group of platelet counts. The line and bar represent the median and IQR for the corresponding index.

3.4. Correlation of platelets with platelets related markers and viral load

To further explore the factors influencing low platelet counts, we analyzed the dynamics of blood parameters associated with platelet levels, including mean platelet volume (MPV), large platelet ratio (P-LCR), platelet distribution width (PDW), and platelet hematocrit (PCT). Similar to the trend observed for HCT, MPV exhibited a gradual increase during the first 7 days after onset, followed by a slow decline, although it remained within the normal range throughout the observation period (Fig. 4A). Comparable patterns were observed for P-LCR and PDW (Fig. 4B–C). However, the dynamics of PCT showed a rapid decline after disease onset, with a trend similar to that of platelets, both reaching their lowest levels on day 6 before increasing rapidly (Fig. 4D), suggesting a potential correlation between these two parameters. Indeed, Pearson correlation analysis revealed that platelet counts were negatively correlated with MPV (Fig. 4A, R = −0.55, p < 2.2e-16), P-LCR percentage (Fig. 4B, R = −0.57, p < 2.2e-16), and PDW percentage (Fig. 4C, R = −0.5, p < 2.2e-16), while showing a strong positive correlation with PCT percentage (Fig. 4D, R = 0.97, p < 2.2e-16). Furthermore, Pearson correlation analysis also showed a negative correlation between platelet count and HCT (Fig. 4E, R = −0.25, p = 2.1e-05), as well as viral load upon admission (Fig. 4F, R = 0.25, p = 0.028), i.e., the lower the NS5 Ct values, the lower the extreme platelet count, indicating that high DENV viral load may contribute to platelet reduction in patients. Patients with higher viral loads (CT < 25, n = 46) had a significantly higher risk of developing severe thrombocytopenia compared to those with CT ≥ 25 (n = 34) (26.1 % vs 8.8 %, RR = 3.25, 95 %CI 1.07–9.84),with a number needed to harm (NNH) was 5.1. Additionally, compared to the control group, patients in the low platelet, intermediate-low platelet, and very low platelet groups exhibited significantly higher percentages of PDW, MPV, and P-LCR, while PCT percentages were significantly reduced. Baseline characteristics of dengue patients with and without poor platelet decline are presented in Table 1.

Fig. 4.

Dynamics and correlation analysis of platelet-related factors. (A–D) Dynamics of mean platelet volume, large platelets ratio, platelet distribution width percentage, and platelet hematocrit, and their correction with platelet counts. (E) Correction between hematocrit and platelet counts. (F) Correction between viral load on admission and platelet counts. Abbreviations: MPV, mean platelet volume; P-LCR, large platelets ratio; PDW, platelet distribution width; PCT, platelet hematocrit.

Table 1.

Baseline characteristics of dengue patients with and without poor platelet decline.

Patients with decreased PLT count
Variable	Very low	Intermediate-low	low	normal	P value (Normal vs very low)	P value (Normal vs Inter-low)	P value (Normal vs low)
Sex n(%)
male	15 (78.9 %)	41 (64.1 %)	16 (69.6 %)	16 (72.7 %)
female	4 (21.1 %)	23 (35.9 %)	7 (30.4 %)	6 (27.3 %)	0.8587	0.1782	0.9475
Age
Mean(SD)	48.26(14.60)	41.48(16.17)	40.57(15.58)	33.96(17.93)	0.0085	0.0706	0.1933
Median(IQR)	51(37, 57)	37(28.25, 54)	39(31, 52)	31(24, 44.75)
Min,Max	19, 88	10, 79	7, 73	1.83, 67
WBC
Mean(SD)	5.03(2.243)	9.882(14.44)	3.190(1.034)	4.536(3.114)	0.5694	0.2251	0.0561
Median(IQR)	3.88(3.37, 6.71)	3.195(2.413, 9.05)	2.97(2.63, 3.77)	3.345(2.558, 5.238)
Min,Max	2.63, 10.4	1.16, 63	1.48, 6.68	1.63, 13.23
HCT
Mean(SD)	45.36(3.637)	42.16(3.632)	42.1(4.922)	42.69(5.752)	0.089	0.6213	0.7127
Median(IQR)	45.8(42.1, 47.9)	42.1(39.3, 44.45)	42.9(37.9, 44.9)	42.8(39.05, 47.35)
Min,Max	39.2, 52.3	34.7, 50.8	31.6, 50.7	29.3, 51
PDW
Mean(SD)	15.68(3.008)	14.31(2.032)	13.3(2.13)	11.95(1.893)	<0.0001	<0.0001	0.0343
Median(IQR)	15.8(13.3, 18.3)	14.1(12.8, 15.9)	13.25(11.3, 14.8)	12.2(10.35, 13.2)
Min,Max	9.7, 21.5	9.6, 19.2	10.2, 17.2	8.1, 15.5
MPV
Mean(SD)	12.26(0.9783)	11.48(0.8586)	10.94(0.9787)	10.42(1.015)	<0.0001	<0.0001	0.0966
Median(IQR)	12.4(11.6, 12.9)	11.5(10.9, 11.9)	10.8(9.1, 11.8)	10.6(9.65, 10.9)
Min,Max	10, 13.9	9.4, 13.5	9.1, 13	8.5, 12.4
PCT
Mean(SD)	0.05316(0.01945)	0.08492(0.01635)	0.1218(0.01868)	0.169(0.03714)	<0.0001	<0.0001	<0.0001
Median(IQR)	0.05(0.05, 0.06)	0.08(0.07, 0.1)	0.12(0.11, 0.13)	0.16(0.145, 0.175)
Min,Max	0.01, 0.1	0.05, 0.12	0.1, 0.18	0.12, 0.26
P-LCR
Mean(SD)	42.81(7.517)	36.94(6.703)	32.72(7.386)	27.9(7.597)	<0.0001	<0.0001	0.0411
Median(IQR)	43.8(36.8, 47.9)	37.1(19.5, 41.5)	31.55(27.85, 38.6)	28.8(21.85, 31.85)
Min,Max	24.8, 56.4	19.5, 50.2	17.6, 49.2	14.1, 43.2

4. Discussion

Over the past decade, the global burden of dengue illness has escalated, with significant outbreaks occurring in endemic regions [3]. Elevated levels of AST, ALT, LDH, ferritin, and creatinine have been identified as risk factors associated with mortality, alongside the lowest platelet counts, hepatomegaly, severe organ involvement, severe bleeding, severe plasma leakage, advanced age at hospitalization, and elevated APACHE II, SAPS II, and SOFA scores [15]. Clinicians are understandably concerned about inadequate platelet recovery; however, data regarding the kinetics of platelet responses in dengue remain limited [11]. In this study, we conducted a systematic analysis of hemocyte dynamics using the complete cohort from our hospital. Notably, our investigation provides a comprehensive and clear depiction of platelet responses throughout the course of dengue infection in the host for the first time. Contrary to previous studies that concentrated on severe thrombocytopenia (<20 × 10⁹/L) in dengue, our findings emphasize the predominance of moderate platelet reduction (50–99 × 10⁹/L) and its correlation with viral load, suggesting a paradigm shift in the monitoring of platelet thresholds.

WHO guidelines previously classified symptomatic dengue virus infections as dengue fever, dengue haemorrhagic fever (DHF), or dengue shock syndrome (DSS). However, the revised WHO guidelines in 2009 classified symptomatic dengue as severe dengue and non-severe dengue, the later is further divided into two subcategories: dengue without warning signs, and dengue with warning signs [9]. An increased HCT concurrent with a rapid decline in platelet count, or thrombocytopenia, is one of the warning signs in dengue infection, and thrombocytopenia serves as a critical indicator of potential severe manifestations [16]. Due to this association, researchers have suggested using thrombocytopenia as a potential predictor of dengue disease progression [17], and numerous studies have investigated predictive factors and clinical outcomes related to poor platelet recovery, as well as the efficacy of prophylactic platelet transfusion in correcting low platelet counts [11,17,18]. Current research indicates that dengue patients with thrombocytopenia who are older, presented earlier in the disease course, or have lower white blood cell counts are more likely to experience poor platelet recovery [11]. This suggests that the potential for irreversible consequences of thrombocytopenia should be evaluated in conjunction with patient-specific factors such as age, underlying diseases, and immune status—including parameters like white blood cell and lymphocyte percentages. Previous studies have reported a correlation between severe dengue and lymphopenia [19], but not neutropenia [20]. Acute dengue infection can lead to both direct and indirect lymphocyte apoptosis through viral infection and the release of proinflammatory cytokines [21,22]. Our hemodynamic observations revealed that during the acute phase of dengue virus infection phase, with the decrease of platelets at 3–9 days onset, the percentage of white blood cells and neutrophils also decreased, while the percentage of lymphocytes increased, and the percentage of lymphocytes in the platelet-decline phase was significantly lower than that in the pre-decline phase. These findings suggest that the reduction in platelets during the acute stage of dengue infection may partially impair the protective inflammatory response mediated by immune cells. Further investigations are underway to determine the roles of inflammatory and vascular markers during the platelet-decline phase of dengue.

Thrombocytopenia is a common symptom in DENV infection, and the threshold used for defining thrombocytopenia is 150 or 100 × 10⁹/L [23]. Similar to the findings of Khatri et al. [24] in an Indian cohort, moderate thrombocytopenia was the most common phenotype in our study (50-99 × 10⁹/L: 50 % vs. 20-99 × 10⁹/L: 64.3 % in India). Notably, neither study reported bleeding complications despite the presence of severe thrombocytopenia (<20 × 10⁹/L: 16.9 % in India; <50 × 10⁹/L: 14.8 % in our study). While Khatri et al. observed no significant changes in mean platelet volume (MPV: 9.01 fL) or platelet distribution width (PDW: 17.2 %) in thrombocytopenic patients, our study found significant inverse correlations between platelet counts and MPV (R = −0.55) as well as platelet-large cell ratio (P-LCR; R = −0.57). These differences may reflect variations in study design (cross-sectional vs. longitudinal) or population characteristics. In addition, our study also found that the abnormal reduction of platelets was independent of the patient's age, gender, and whether the patient had underlying diseases, and there was no difference in the number of platelets between elderly patients and patients of other ages, so as it between underlying health conditions (such as chronic liver diseases, diabetes, hypertension, pregnant women) and healthy population.

While dengue-associated thrombocytopenia is well-documented, its pathophysiology involves both established and hypothesized mechanisms. Dengue-induced thrombocytopenia is postulated to result from peripheral platelet destruction, platelet sequestration, and virus-induced bone marrow suppression [25,26]. Strong experimental evidence demonstrates that DENV directly infects megakaryocytes, reducing CD41/CD61 expression and causing mitochondrial dysfunction [27], while immune-mediated destruction is driven by NS1-elicited antibodies that promote platelet phagocytosis [28,29]. Clinical observations further confirm that proinflammatory cytokines (IL-1β, TNF-α, IFN-γ) suppress megakaryopoiesis and accelerate platelet clearance [29]. Potential but less-verified mechanisms include complement activation (associated with elevated C3a/C5a in severe dengue [26,29]) and platelet apoptosis (suggested by annexin V/P-selectin markers [30,31]). Our findings support this multifactorial model: the inverse correlation between platelet count and viral load (R = −0.25), along with altered platelet indices (increased MPV and decreased PCT), reflects a combination of peripheral platelet consumption and compensatory bone marrow response. However, the precise roles of complement and apoptotic pathways warrant direct experimental validation in patients with dengue.

The two IgG-positive cases in our cohort both developed very low platelet counts (40 and 49 × 10⁹/L) during the critical phase, falling within the ≤49 × 10⁹/L severity stratum (14.8 % of total cohort). While this observation aligns with reports of enhanced thrombocytopenia in secondary infections [32], the limited sample size prevents a statistically meaningful comparison with IgG-negative cases. Importantly, none of the IgG-positive cases progressed to severe dengue, despite fulfilling the WHO platelet criteria for severe thrombocytopenia—a trend that reflects the overall pattern observed in our cohort. These findings suggest that, during this DENV-2-dominant epidemic, IgG positivity may not be a reliable predictor of severe outcomes in homotypic reinfections, and that platelet thresholds for defining severe dengue may need to be calibrated according to serostatus. Further validation in cohorts with confirmed heterotypic reinfections and quantitative IgG assessments is warranted.

Our findings carry significant implications for public health strategies in dengue-endemic regions, particularly in resource-limited settings. The biphasic platelet decline pattern (nadir at day 6, recovery by day 9) provides a critical time window for targeted monitoring. In high-transmission urban areas like Shenzhen, integrating daily platelet count assessments during days 3–9 of illness could optimize triage systems, reducing unnecessary hospitalizations for patients with stable platelet levels (≥50 × 10⁹/L) and prioritizing care for those at risk of severe thrombocytopenia. Furthermore, the strong correlation between platelet counts and viral load (Ct values) suggests that point-of-care NS5 RNA quantification could serve as a cost-effective prognostic tool in primary healthcare centers, where advanced laboratory infrastructure is often lacking. These approaches align with the WHO's 2023 roadmap for dengue control, which emphasizes context-specific biomarkers to alleviate healthcare burdens during outbreaks.

While this study provides novel insights into platelet dynamics during dengue infection, several limitations should be acknowledged. First, the single-center design limits generalizability to populations with different genetic backgrounds or viral genotypes; however, Consistent with Shenzhen's endemic dengue profile (Introduction Section), our findings reflect hematological manifestations in an urban transmission setting. Second, the moderate sample size (n = 135) allowed robust overall analyses but constrained subgroup comparisons (e.g., pediatric or elderly cohorts), warranting validation in larger, age-diverse cohorts. Third, retrospective data spanning 2014–2023 were included, yet standardized laboratory protocols and consistent use of the same hematology analyzer minimized potential biases from temporal variations. Fourth, viral load quantification relied on NS5 Ct values, which, though practical for clinical settings, may not fully capture infectivity or host-pathogen interactions; future studies integrating viral culture or sequencing could enhance mechanistic insights. Fifth, the absence of severe dengue cases in this cohort precluded direct exploration of thrombocytopenia-associated complications (e.g., hemorrhage), but this also reflects the success of early clinical interventions in preventing progression. In addition, while we recorded baseline comorbidities (chronic liver diseases/diabetes/hypertension/pregnant women prevalence: 12.3 % [34/135]) and hospitalization treatments (i,e. NSAIDs), our study was not powered to adjust for all potential confounders. The observed platelet dynamics should therefore be interpreted as real-world associations rather than isolated causal relationships, particularly given dengue's multifactorial pathophysiology. Moreover, serum samples from 62/135 patients (45.9 %) collected within 7 days of symptom onset were tested for dengue-specific IgM and IgG antibodies using ELISA (Wantai Biological). While qualitative results (positive/negative) were available, quantitative titers and avidity testing were not routinely performed due to clinical laboratory constraints. All cases were confirmed as DENV-2 infections by NS5 gene RT-PCR, but secondary serotype infections could not be excluded without convalescent-phase serology. Finally, unmeasured confounders (e.g., genetic polymorphisms, detailed cytokine profiles) were not analyzed, which could refine our understanding of platelet-virus crosstalk. Despite these limitations, our prospective-longitudinal approach and integration of multi-year data offer actionable evidence for thrombocytopenia monitoring, emphasizing the need for context-specific platelet thresholds in dengue management.

5. Conclusion

This study demonstrates that platelet dynamics in dengue fever exhibit characteristic biphasic reduction, with nadir levels (mean 97.65 × 10⁹/L) detectable at day 6 post-onset and complete recovery by day 9, supporting the clinical utility of platelet monitoring during this critical window. The absence of severe dengue progression despite 14.8 % patients developing severe thrombocytopenia (<50 × 10⁹/L) suggests that conventional platelet thresholds may overestimate bleeding risk in this population. The strong correlations between platelet counts and both viral load (R = 0.25, p = 0.028) and lymphocyte/neutrophil ratios (p < 0.001) provide mechanistic insights into dengue-associated thrombocytopenia. Future studies should establish standardized platelet value thresholds integrated with virological and immunological parameters, which is essential for risk stratification. If validated, this multiparameter approach could differentiate between viral-mediated platelet destruction and immune dysregulation, ultimately guiding personalized management and improving dengue outcomes.

CRediT authorship contribution statement

Liping Guo: Writing – original draft, Funding acquisition, Conceptualization. Yuchen Gu: Writing – original draft, Formal analysis, Data curation. Ying Zhang: Methodology, Investigation, Data curation. Haimei Zhang: Software, Methodology. Weizhen Weng: Resources, Data curation. Shuai Wu: Methodology, Investigation. Jing Yuan: Writing – review & editing, Funding acquisition.

Data availability statement

The data for this study is available upon reasonable request.

Funding

This work was supported by Shenzhen Fund for Guangdong Provincial High-Level Clinical Key Specialties (No. SZGSP011), Shenzhen Medical Research Fund (B2402042), First-class discipline innovation-driven talent program of Guangxi Medical University and Key Laboratory of Basic Research on Regional Diseases (Guangxi Medical University), Education Department of Guangxi Zhuang Autonomous Region. All authors critically reviewed the manuscript, gave their final approvals, and are accountable for accuracy and integrity.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: