Background: Certain biochemical markers have been found useful in monitoring the progress and predicting the severity of dengue. Serum lactate dehydrogenase (LDH) is one such potential biomarker. Aim and objective: To study the serum LDH levels in critically ill patients of dengue with warning signs and severe dengue and its relationship with severity of illness. Materials and methods: Retrospective analytical study. Results: Sixty patients with severe dengue and dengue with warning signs were included. Mean duration of fever at admission was 4 ± 1.6 days and the mean age was 9.6 ± 5.06 years. Male:female ratio was 1:1. Half the children were above 10 years. The median LDH value was 1,133.5 IU [interquartile range (IQR) 640–1,732]. The mean LDH value was significantly higher in patients with severe dengue (2,986.65 ± 3,638.54) as compared to dengue with warning signs (1,209.87 ± 1,370.20) (p—0.047). The majority (70%) of patients with severe dengue had LDH >1,000 IU and complications like severe bleeding, pleural effusion, ARDS, and shock were higher in this group. Mean hospital stay in patients with LDH >1,000 was 14.685 ± 5.993 days and in those with LDH <1,000 was 8.732 ± 3.312 days (p = 0.000). Mean platelet count was significantly lower in severe dengue (56,405.00 ± 49,918.74) as compared to dengue with warning signs (922,257.50 ± 71,235.44) (p value 0.028) and there was a weak negative correlation between LDH and platelet count which was non-significant (r = Karl Pearson coefficient −0.055; p value 0.676). The case fatality rate was 9%. The mean LDH (4,783 ± 5,131) in non-survivors was much higher than survivors (1,531.1 ± 1,986) though this was not statistically significant. Conclusion: Serum LDH level at admission was significantly raised in severe dengue as compared to dengue with warning signs. Similarly, mean LDH values were higher in survivors as compared to non-survivors. There was a weak negative correlation between LDH and platelet count. Clinical significance: Lactate dehydrogenase can be used early in the disease to identify those who may progress to severe dengue and predict mortality. This will help optimize resource allocation and more effective care in a disease-endemic country like India.