Table 3

Multivariate analysis of the association between baseline covariates and in-hospital mortality in 303 patients with confirmed bacterial meningitis using logistic regression modelling

Baseline covariate	N	In-hospital mortality N (%)*	Crude OR for in-hospital mortality (95% CI)	P value	Adjusted OR for in-hospital mortality (95% CI)†	P value‡
Sex
Male	173	26 (15.1)	1
Female	130	12 (9.23)	0.57 (0.27 to 1.18)	0.13
Age group
≤18 years	18	0 (0)
19–59 years	159	18 (11.3)	1
≥60 years	126	20 (16.0)	1.49 (0.75 to 2.96)	0.25
Blood culture positive
No	137	11 (8.09)	1		1
Yes	166	27 (16.3)	2.21 (1.04 to 4.67)	0.03	1.87 (0.87 to 4.01)	0.10
GCS≤13§
No	124	8 (6.45)	1		1
Yes	148	27 (18.2)	3.24 (1.39 to 7.52)	0.004	2.90 (1.26 to 6.71)	0.008
IV dexamethasone given¶
No	149	23 (15.4)	1		1
Yes	150	14 (9.40)	0.57 (0.27 to 1.16)	0.11	0.57 (0.28 to 1.17)	0.12
Intravenous dexamethasone given if Streptococcus pneumoniae**
No	73	16 (21.9)	1		1
Yes	97	11 (11.5)	0.46 (0.20 to 1.08)	0.07	0.47 (0.20 to 1.10)	0.08
Final diagnosis S. pneumoniae
No	131	10 (7.63)	1		1
Yes	172	28 (16.4)	2.37 (1.10 to 5.11)	0.02	2.08 (0.96 to 4.48)	0.05
ICU admission††
No	144	7 (4.86)	1		1
Yes	157	31 (19.7)	4.81 (1.99 to 11.60)	<0.001	4.28 (1.81 to 10.1)	<0.001

*7/11 patient with progressing rash, 131/176 patients with GCS <13 and 13/16 patients with uncontrolled seizures.
†Adjusted for sex and age group.
‡P value from Likelihood ratio test comparing models with and without primary exposure variable.
§31/303 (10%) participants did not have a GCS recorded.
¶4/303 (1%) participants had missing data on intravenous dexamethasone administration.
**2/172 (1%) participants with confirmed S. pneumoniae meningitis had missing data on intravenous dexamethasone administration.
††1/303 (0.3%) participants had missing data on ICU admission.
GCS, Glasgow Coma Score.