Table 6.
Univariate (U) or multivariate (M) regression results for clinical outcomes by hydrocephalus aetiology
| Outcome | Odds ratio (95% confidence intervals)/median (range) | |||||
|---|---|---|---|---|---|---|
| PHH | NTDs | Genetic | Congenital | Other | p value | |
| CP (M) |
n = 85 1 (reference) |
n = 34 0.022 (0.002–0.206)* |
n = 26 0.310 (0.083–1.158) |
n = 38 0.221 (0.059–0.820) * |
n = 54 0.338 (0.101–1.128) |
0.01 |
| Epilepsy (M) |
n = 94 1 (reference) |
n = 36 0.133 (0.030–0.581)* |
n = 29 1.320 (0.384–4.537) |
n = 54 0.630 (0.199–1.993) |
n = 67 0.383 (1.114–1.281) |
< 0.01 |
| Speech delay (M) |
n = 93 1 (reference) |
n = 29 0.648 (0.234–1.792) |
n = 38 4.230 (0.914–19.572) |
n = 52 0.272 (0.125–0.592)* |
n = 63 0.240 (0.114–0.502)* |
< 0.01 |
| Schooling (U) | Mainstream school with support (reference: mainstream school) | |||||
|
n = 50 1 (reference) |
n = 23 1.556 (0.447–5.413) |
n = 25 2.000 (0.334–11.969) |
n = 38 0.407 (0.136–1.220) |
n = 50 0.293 (0.101–0.854)* |
< 0.01 | |
| Special school (reference: mainstream school) | ||||||
|
n = 50 1 (reference) |
n = 23 0.200 (0.042–0.959)* |
n = 25 3.400 (0.661–17.500) |
n = 38 0.200 (0.068–0.592)* |
n = 50 0.224 (0.084–0.598)* |
< 0.01 | |
| Behavioural disorders (M) |
n = 88 1 (reference) |
n = 35 0.871 (0.220–3.449) |
n = 29 1.572 (0.416–5.939) |
n = 53 2.601 (0.799–8.468) |
n = 64 2.047 (0.594–7.051) |
0.23 |
| Endocrine dysfunction (U) |
n = 87 1 (reference) |
n = 35 0.693 (0.137–3.510) |
n = 34 4.114 (1.390–12.176)* |
n = 53 1.190 (0.358–3.961) |
n = 64 1.404 (0.467–4.222) |
0.06 |
| Mortality (M) |
n = 96 1 (reference) |
n = 35 |
n = 41 6.070 (1.137–32.390)* |
n = 60 1.771 (0.306–10.248) |
n = 70 0.932 (0.131–6.630) |
0.01 |
| Age at death (years)˟ |
n = 7 1.91 (0.70–5.18) |
n = 0 |
n = 12 1.16 (0.02–5.62) |
n = 7 1.42 (0.81–5.63) |
n = 4 1.93 (0.31–4.69) |
0.69 |
Predictors included in the final model varied by clinical outcome are as follows:
CP: birthweight (a priori — gestational age excluded due to high multicollinearity) and number of revisions < 2 years of age (significant in model: significantly increased the odds of CP by 34.1% for every revision, p = 0.04)
Epilepsy: gestational age (a priori), corrected age at VP shunt insertion (a priori), shunt revision ≤ 1 year of age (a priori), number of revisions < 1 year of age (a priori), and shunt infection ≤ 1 year of age (a priori)
Speech delay: sex (significant in model: male sex significantly increased odds of speech delay by 49.5%, p = 0.02)
Schooling: there is no multivariate model for schooling as it was not significantly affected by any of the possible predictors included in model building, nor were any predictors indicated for a priori inclusion
Mortality: gestational age (a priori) and corrected age at VP shunt insertion (a priori and significant in model: odds of mortality decreased by 0.6% for every week older the infant was at shunt insertion, p = 0.04)
Behavioural outcome: gestational age (a priori), number of revisions before age 2 (a priori), and sex (a priori and significant in model: male sex significantly increased odds of behavioural disorders by 43.3%, p < 0.05)
Endocrine dysfunction: there is no multivariate model as the outcome was not significantly affected by any of the possible predictors, nor were any predictors indicated for a priori inclusion
PHH post-haemorrhagic hydrocephalus, NTDs neural tube defects, CP cerebral palsy
˟Median (range)
*Parameters that are statistically significant between that aetiology group and PHH (using logistic regression)