Patients’ demographics characteristics
Among the 325 enrolled patients with laboratory-confirmed dengue, 130 (40.0%) were classified as having dengue with warning signs (DWS) and 195 (60.0%) as dengue without warning signs (DWWS) based on the WHO 2009 classification. Notably, no cases of severe dengue were identified among the enrolled participants. Table 1 presents the sociodemographic distribution of patients stratified by dengue severity. The median age of patients in the DWS group was 40 years (IQR 29–52), significantly higher than the 33 years (IQR 21–48) observed in the DWWS group. However, age group distribution did not differ significantly between the two severity groups (p = 0.9168), with children, adults, and elderly individuals represented similarly in both groups.
A significant gender disparity was noted, with females comprising 69.2% of the DWWS group compared to 30.8% in the DWS group (p = 0.0004). Among adults aged ≥ 20 years, BMI distribution showed no significant association with disease severity (p = 0.2936), although overweight and obese individuals were slightly more prevalent in the DWS group (Table 1).
Liver enzyme elevation and disease severity
We further examined the association between liver enzyme elevation and disease severity among the 325 dengue patients analyzed. AST elevation was found more common than ALT elevation. AST levels showed marked stratification by disease severity, with 25.5% of DWS patients exhibiting > 1–3× upper limit of normal (ULN) elevations versus 29.5% in DWWS. Notably, 3.1% of DWS patients demonstrated > 3–10× ULN AST elevations (vs. 2.5% DWWS), while no cases exceeded 10× ULN. ALT elevations were less pronounced, with only 5.5% of DWS patients showing > 1–3× ULN increases and no DWS cases exceeding 3× ULN (Fig. 1).
Distribution of liver enzyme elevation by dengue severity
Violin plots (Fig. 2) demonstrate significantly elevated AST levels in DWS patients (median: 61.5 U/L, IQR: 36.3–90.9) compared to DWWS (median: 44.0 U/L, IQR: 24.6–75.8), with wider dispersion and a pronounced right-skewed density peak above 80 U/L. While ALT medians were comparable between groups (DWS: 27.1 U/L vs. DWWS: 28.2 U/L), DWWS patients exhibited greater variability including an outlier at 182.0 U/L (4.6× ULN) (Fig. 2).
Violin plots of liver enzyme distributions by dengue severity
The Chi square analysis revealed a significant association between elevated AST levels and dengue with warning signs (DWS). Specifically, patients with elevated AST were more likely to be in the DWS group compared to those with normal AST levels (48.9% vs. 29.8%), with a statistically significant p-value of 0.0006 and an odds ratio (OR) of 0.44 (95% CI: 0.28–0.70). This suggests that elevated AST may serve as a notable biomarker for more severe disease manifestations. In contrast, ALT levels showed a non-significant trend toward association with DWS, with a p-value of 0.1809 and an OR of 1.56 (95% CI: 0.85–2.86), indicating ALT may be less specific for disease severity in this context (Table 2).
Clinical characteristics by dengue severity
To establish the clinical context of dengue severity, we analyzed key symptoms among the 325 patients, stratified by DWWS (n = 195, 60.0%) and DWS (n = 130, 40.0%), as presented in Table 3. Fever was nearly universal (99.1%), with no significant difference between severity groups (p = 1.000). Gastrointestinal symptoms, particularly nausea (33.8% in DWS vs. 14.9% in DWWS, p < 0.0001), vomiting (21.5% vs. 7.7%, p = 0.0004), and abdominal pain (16.9% vs. 0.0%, p < 0.0001), were significantly more prevalent in DWS, suggesting a link to severe disease manifestations. Notably, joint pain emerged as another discriminating feature, present in nearly half of DWS patients (47.7%) versus 34.4% in DWWS (p = 0.021), suggesting systemic inflammatory responses may parallel hepatic injury (Table 3).
Clinical characteristics by liver enzyme elevation status
Clinical features were analyzed in relation to liver enzyme status to identify symptoms associated with elevated AST or ALT levels (Table 4). Patients with elevated AST demonstrated significantly higher rates of gastrointestinal symptoms, including nausea (33.3% vs. 9.9%, p < 0.0001) and vomiting (18.4% vs. 7.3%, p = 0.003), aligning with our earlier observation that these symptoms cluster in severe dengue cases (Table 3). Unexpectedly, malaise was 2× less prevalent in elevated AST patients (23.0% vs. 45.0%, p < 0.0001), suggesting potential pathophysiological differences in energy metabolism during hepatic injury. 100% of patients with elevated AST/ALT reported fever, reinforcing fever as a universal marker of dengue-associated liver injury. Rash was less frequent in patients with elevated AST (p = 0.0289). For ALT abnormalities, a paradoxical pattern emerged: elevated ALT was associated with absence of abdominal pain (0% vs. 8.2%, p = 0.019) and reduced headache prevalence (33.3% vs. 58.2%, p < 0.0001) (Table 4).
These findings reinforce the clinical relevance of liver enzyme monitoring, particularly AST, in the assessment of dengue patients, not only as a marker of hepatic involvement but also in predicting symptom patterns linked to more severe or systemic disease. The clinical symptom burden appears more pronounced among individuals with elevated AST than ALT, aligning with earlier findings that AST had a stronger association with disease severity (Table 2).
Diagnostic and treatment profiles
In alignment with earlier findings that associated AST elevation with dengue severity, Table 5 reveals critical diagnostic and management differences among dengue patients with liver injury. Elevated AST showed a striking association with NS1 antigen positivity (83.3% vs. 44.4%, p = < 0.0001), suggesting enhanced early viral detection in hepatic-involved cases, a finding that aligns with our earlier observation of more frequent vomiting and fever in this group (Table 4). Conversely, IgM positivity was paradoxically lower in AST-elevated patients (18.4% vs. 63.6%, p = < 0.0001), potentially reflecting delayed seroconversion due to immune modulation during severe hepatic injury.
The need for intravenous (IV) fluid support was significantly greater among those with elevated AST (9.8% vs. 0.7%, p = 0.0003), consistent with a more severe clinical presentation and previously noted associations between elevated AST and gastrointestinal symptoms such as nausea and vomiting (Table 4). No statistically significant differences were observed in treatment or diagnostic patterns when stratified by ALT status (Table 5).
Adjusted associations and exploratory analyses
In multivariable logistic regression analysis, elevated AST was significantly associated with increased odds of developing dengue with warning signs (DWS), with an adjusted odds ratio (aOR) of 2.40 (95% CI: 1.49–3.86, p = 0.0003). Female gender was also independently associated with higher odds of DWS (aOR = 2.31, 95% CI: 1.45–3.70, p = 0.0005). Age demonstrated a modest but statistically significant association with severity (aOR = 1.01 per year increase, 95% CI: 1.00–1.03, p = 0.0299). Although elevated ALT showed a trend toward a protective effect, the association did not reach statistical significance (aOR = 0.57, 95% CI: 0.30–1.09, p = 0.0900). The logistic regression model demonstrated an acceptable fit, with a McFadden’s pseudo R² of 0.074 and an AIC of 414.9 (Table 6). Multicollinearity diagnostics using Variance Inflation Factor (VIF) indicated no concerning collinearity among predictors (VIFs ranged from 1.003 to 1.014 for all variables), confirming the model’s internal consistency and robustness.
Spearman’s correlation analysis was conducted to explore the relationships between liver enzyme levels (AST and ALT) and various clinical and laboratory parameters (Table 7). In patients with elevated liver enzymes, AST levels in DWS patients were predominantly > 1–3× the ULN (48 U/L, 25.5%) with 3.1% reaching > 3–10× ULN, while ALT elevations were less pronounced, with 5.5% of DWS patients at > 1–3× ULN (55 U/L) and none exceeding 3× ULN (Fig. 1). The selected parameters were chosen based on their clinical relevance and previously documented associations with dengue severity or hepatic involvement. For instance, WBC count, platelet count, and hemoglobin are commonly monitored in dengue management due to their relevance in assessing disease progression. Serum creatinine and hematocrit were included to evaluate renal function and hydration status, respectively. Clinical features such as headache, malaise, nausea, vomiting, and abdominal pain were included as they are frequently reported symptoms in dengue patients and may correlate with systemic inflammation or hepatic dysfunction. The analysis revealed several significant associations.
AST showed strong positive correlations with hemoglobin (ρ = 0.407, p < 0.0001) and platelet count (ρ = 0.277, p < 0.0001), suggesting potential interactions between liver function and hematological parameters. AST also correlated with nausea (ρ = 0.243, p < 0.0001) and vomiting (ρ = 0.151, p = 0.0064), linking elevated AST to gastrointestinal symptoms (Table 7).
ALT demonstrated a significant negative correlation with headache (ρ = −0.177, p = 0.0013) and a weaker negative correlation with malaise (ρ = −0.149, p = 0.0072). No significant association was found between ALT and hemoglobin (ρ = −0.003, p = 0.9577) or platelet count (ρ = 0.020, p = 0.7245) (Table 7).
Both enzymes showed weak negative correlations with WBC count and serum creatinine, though these associations were only marginally significant for AST. These findings highlight distinct patterns of correlation for AST and ALT with clinical and laboratory parameters, providing further insight into the pathophysiology of dengue-associated liver injury.