Table 2.
Statistical comparisons of CSF results, clinical and laboratory data
| (log) AI | Tau | GFAP | S100b | NSE | |
|---|---|---|---|---|---|
| Demography and general data | |||||
| Age | 0.0215 | 0.0148 | NS | NS | NS |
| Weight | NS | NS | NS | NS | NS |
| Clinical signs and symptoms | |||||
| Headache | NS | NS | NS | NS | NS |
| Vomiting | NS | NS | NS | NS | NS |
| Seizures | 0.0055 | 0.0003 | NS | NS | NS |
| Rash | NS | NS | NS | NS | NS |
| Hearing loss | NS | 0.0075 | NS | NS | NS |
| Photophobia | NS | NS | NS | NS | NS |
| Eschar | NS | NS | NS | NS | NS |
| Visual loss | NS | 0.0422 | NS | NS | NS |
| Severity and outcome measures | |||||
| Outcome | NS | NS | NS | NS | NS |
| GCS | NS | 0.0095 | 0.0107 | 0.0237 | 0.0213 |
| WHO meningism | NS | 0.0134 | 0.0272 | 0.0064 | 0.0427 |
| WHO AES | NS | 0.0001 | 0.0312 | 0.0126 | NS |
| WHO men and AES | NS | 0.0007 | NS | 0.0087 | NS |
| Laboratory investigations | |||||
| CSF opening pressure | NS | NS | NS | NS | NS |
| Turbidity | NS | NS | NS | NS | NS |
| CSF white cells/mm3 | < 0.0001 | NS | 0.0038 | 0.0024 | 0.0030 |
| CSF neutrophils/mm3 | 0.0001 | NS | 0.0012 | 0.0017 | 0.0036 |
| CSF lymphocytes/mm3 | 0.0004 | NS | NS | 0.0337 | 0.0255 |
| Blood/CSF glucose ratio | 0.0001 | NS | NS | NS | 0.0437 |
| CSF lactate > 4 mmol/L | 0.0001 | NS | 0.0001 | 0.0446 | 0.0143 |
| CSF glucose < 2.5 mmol/L | NS | NS | NS | NS | NS |
| CSF protein > 40 mg/L | 0.0020 | NS | NS | 0.0481 | 0.0303 |
| Bilirubin | NS | NS | NS | NS | NS |
| Hematocrit | NS | NS | NS | NS | NS |
AES = acute encephalitic syndrome; AI = albumin index; CSF = cerebrospinal fluid; GFAP = glial fibrillary acidic protein; NS = nonsignificant; NSE = neuron-specific enolase; WHO = World Health Organization.
Comparisons across clinical groups were made using the Kruskal–Wallis equality-of-populations rank test. Because of the exploratory nature of this study and multiple comparisons, a conservative P value of < 0.01 was considered significant (shown in bold). Exact P values are reported (for values < 0.05) for Bonferroni correction (α/n, where α = 0.05 and n = number of tests), if preferred.