Clinical factors predicting rotavirus diarrhea in children: A cross-sectional study from two hospitals

BACKGROUND Rotavirus is still a significant contributing morbidity and mortality in pediatric patients. AIM To look at clinical signs and symptoms and laboratory findings that can predict rotavirus gastroenteritis compared to non-rotavirus gastroenteritis. METHODS This was a cross-sectional study with medical records obtained from December 2015 to December 2019. Inclusion criteria for this study include all hospitalised pediatric patients (0-18 years old) diagnosed with suspected rotavirus diarrhea. The receiver operating curve and Hosmer-Lemeshow test would be used to assess the final prediction findings' calibration (goodness of fit) and discrimination performance. RESULTS This study included 267 participants with 187 (70%) rotavirus-diarrhea cases. The patients were primarily male in both rotavirus (65.2%) and non-rotavirus (62.5%) groups. The median age is 1.33 years old (0.08-17.67 years old). Multivariate analysis shows that wet season (ORadj = 2.5; 95%CI: 1.3-4.8, Padj = 0.006), length of stay (LOS) ≥ 3 days (ORadj = 5.1; 95%CI: 1.4-4.8, Padj = 0.015), presence of abdominal pain (ORadj = 3.0; 95%CI: 1.3-6.8, Padj = 0.007), severe dehydration (ORadj = 2.9; 95%CI: 1.1-7.9, Padj = 0.034), abnormal white blood cell counts (ORadj = 2.8; 95%CI: 1.3-6.0, Padj = 0.006), abnormal random blood glucose (ORadj = 2.3; 95%CI: 1.2-4.4, Padj = 0.018) and presence of fecal leukocytes (ORadj = 4.1, 95%CI: 1.7-9.5, Padj = 0.001) are predictors of rotavirus diarrhea. The area under the curve for this model is 0.819 (95%CI: 0.746-0.878, P value < 0.001), which shows that this model has good discrimination. CONCLUSION Wet season, LOS ≥ 3 d, presence of abdominal pain, severe dehydration, abnormal white blood cell counts, abnormal random blood glucose, and presence of fecal leukocytes predict rotavirus diarrhea.