Table 1.
Summary of studies with multivariate analyses in identification of risk factors and detection of CP.
| Study | Subject sample | Data | Methods | Main findings | Other findings |
|---|---|---|---|---|---|
| Pinto-Martin et al. (22) | 113 children with CP | Clinical data (birth weight, gestational age, length of hospital stay, gender, race, plurality, presence of labor, Apgar score, motor function, cranial US findings, etc.) | Multivariate logistic regression to assess risk factors for CP | Risk factors for disabling CP: PEL/VE or ventricular enlargement on cranial US, germinal matrix/intraventricular hemorrhage, mechanical ventilation; risk factors for non-disabling CP: PEL/VE | Cranial US abnormalities are strong risk factors for disabling CP in low birth weight infants; non-risk factors for disabling CP: birth weight, gestational age, length of hospital stay, gender, race, plurality, presence of labor, Apgar score |
| Allan et al. (24) | 36 pts with CP (in 381 infants) | Clinical data (birth weight, bronchopulmonary dysplasia, abnormal cranial US findings, treatment, etc.) | Univariate and multivariate analysis to identify antecedents of CP | Predictors of CP: bronchopulmonary dysplasia and an abnormal cranial US scan (showing grade 3 to 4 intraventricular hemorrhage, PVL, or ventriculomegaly) | PVL and ventriculomegaly associated with high CP detection rates; chorioamnionitis and treatment with surfactant significant in univariate analysis |
| Kim et al. (25) | 35 pts with CP | Clinical data (age, weight, neonatal sepsis, neonatal seizure, etc.) | Univariate and multivariate analysis to identify risk factors for CP | Risk factors for CP and delayed development: neonatal sepsis | |
| Han et al. (27) | 21 children with CP | Clinical data (birth characteristics, disease at birth, neonatal cerebral ultrasound findings, etc.) | Multivariate analysis used to identify risk factors for CP | Risk factors for CP: existence of PVL, preterm labor, preterm rupture of membrane, severe birth asphyxia, neonatal sepsis, and respiratory distress syndrome | Existence of PVL is the strongest risk factor for CP |
| Zhong et al. (28) | 308 children with CP | Data from a cross-sectional survey (birth characteristics, disease during the first month of life, etc.) | Multivariate analysis used to identify risk factors for CP | Risk factors for CP: delivery at home, low Apgar score, illness during the first month of life, maternal cold with fever in early gestation, low protein intake during pregnancy, low education level of mother | |
| Golomb et al. (26) | 76 children with CP after perinatal stroke | Clinical data (perinatal history, motor function, frequency of CP, degree of disability, etc.) | Univariate and multivariate analysis (with logistic regression) to assess risk factors for CP in perinatal stroke | 68% pts with perinatal stroke had CP; risk factors for CP: delayed stroke and male gender; In pts with neonatal stroke, risk factors for triplegia or quadriplegia: bilateral infarcts | In pts with unilateral middle cerebral artery infarcts, risk factors for CP: delayed stroke and large-branch infarction |
| Miamoto et al. (29) | 60 pts with CP vs. 60 healthy controls | Data from questionnaires and clinical exams (TMD symptoms, bio-psychosocial characteristics, etc.) | Multivariate logistic regression to determine risk factors for TMD symptoms | Risk factors for TMD symptoms: presence of CP, male gender, severity of the malocclusion, mouth breathing, and mixed dentition | 13.3% pts vs. 1.7% controls had TMD symptoms |
| Abdullahi et al. (30) | 111 pts with CP vs. 222 controls | Clinical data (maternal sociodemographic characteristics and neonatal expected predictors) | Univariate and multivariate (logistic regression) analyses used to identify factors associated with CP | Predictors of CP: maternal fever, previous neonatal death, and poor sucking | Factors not associated with CP: maternal age, parity, birth weight, and sex |
| Yu et al. (31) | 203 preterm infants with CP, vs. 220 preterm infants without CP or other neurological disorders | Data of diseases of premature infants, the treatments in neonatal period, etc. | Multivariate logistic analysis used to identify risk factors associated with CP | Risk factors for CP: occurrence of PVL, HIE, hypoglycemia, or neonatal jaundice | Continuous positive airway pressure may lower the risk of CP |
| Golomb et al. (32) | 76 children with CP after perinatal stroke | Clinical data (perinatal history, motor function, frequency of CP, degree of disability, etc.) | Univariate and multivariate analysis (with logistic regression) to assess association of CP with other disabilities | 72% pts with perinatal stroke had at least another disability; risk factors for epilepsy: neonatal presentation and history of cesarean-section delivery | Risk factors for severe cognitive impairments or epilepsy: perinatal stroke with neonatal presentation |
| Griffiths et al. (33) | 20 pts with spastic CP; 20 with dyskinetic CP | Injury severity scores at different brain regions on magnetic resonance imaging (T2) | Variables indicated by univariate analysis fed to multivariate logistic regression to identify predictors to differentiate spastic and dyskinetic CP | Spastic CP pts had more severe damage to white matter near the paracentral lobule; dyskinetic CP pts had more injury to the STN: hypoxic-ischemic injury to the STN at birth associated with dyskinetic CP | Non-predictors of dyskinesia: injuries to the putamen, caudate, and globus pallidus |
| Yoshida et al. (34) | 34 pts with CP vs. 21 healthy subjects | Parameters (number of fibers, tract-based FA, and FA) for CST and posterior thalamic radiation tracts from diffusion tensor imaging (DTI) and motor level data | Univariate and multivariate (regression) analysis used to identify variables correlated to gross motor function | Number of fibers and ROI-based FA values of both tracts were lower in pts than controls; motor-sensory parameters were negatively correlated with GMFCS level | |
| Coppola et al. (35) | Group 1: 40 pts with CP and mental retardation; group 2: 47 pts with CP, mental retardation and epilepsy; group 3: 26 pts with epilepsy | Clinical data (age, BMI, BMD z-score from dual-energy X-ray absorptiometry scan, etc.) | Multivariate analysis used to identify factors on BMD | Lower BMD in 42.5% pts in group 1, 70.2% in group 2, 11.5% in group 3 | In pts with CP, mental retardation and epilepsy, epilepsy is an aggravating factor on bone health Factors on BMD: age, BMI, severe mental retardation, epilepsy |
| Benfer et al. (36) | 120 pts with CP | Data of OPD measures, motor measures, etc. | Univariate and multivariate regression analysis to determine the relationship between OPD and motor functions | Higher odds of OPD in non-ambulant pts than in ambulant pts | 85% pts had OPD |
| Romeo et al. (37) | 100 pts with CP (32 of them with epilepsy) vs. 100 healthy children | Data from the SDSC, GMFCS levels, etc. | Multivariate analysis (logistic regression) used to identify factors associated with SDSC | 13% of children with CP had abnormal sleep score; factors associated with SDSC: behavioral problems and epilepsy | Compared with healthy controls, sleep disorders are more common in children with CP |
| Adler et al. (38) | 18 children with unilateral spastic CP (9 with mirror movement, 9 without) | Clinical data from BANIMM, JTHFT, and AHA | Multivariate analysis of covariance used to determine whether mirror movements affect daily living | Mirror movements had a negative impact on bimanual performance (AHA) and on the time needed to complete difficult activities | |
| Tao et al. (39) | 11 children with CP, 8 healthy children, 7 healthy adults | EMG data from five thigh muscles and three lower leg muscles | Multivariate empirical mode decomposition enhanced MMSE analysis used to analyze EMG data; repeated-measure ANOVAs for group comparison | Compared with the control group, CP pts had distinct diversity in MMSE curve | Abnormal MMSE curve reflected problems in individual muscles such as motor control impairments, loss of muscle couplings, and spasticity or paralysis |
| Ghate et al. (40) | 54 pts with CP | Clinical data (CP type, motor function, etc.) and data from ophthalmoscopic examinations | Multivariate logistic regression to identify factors associated with motor outcomes | 70% pts had abnormal optic nerve head; disk pallor associated with non-ambulatory status and quadriplegia; large cup associated with age at examination | Indicator for poor motor outcome: presence of optic nerve head pallor |
| Reid et al. (41) | 31 children with unilateral CP | Activation maps from fMRI (with hand task); FA and MD values and fiber tracts in the thalamocortical and corticomotor tracts from DTI; clinical scores of motor ability | k-means clustering used to identify fMRI-task-specific DTI tracks; surface-based approach (using surface-meshes) compared with voxelwise fMRI-DTI approach; correlation analysis between DTI metrics and clinical scores performed | DTI metrics and five clinical scores of motor function were correlated; surface-based approach processed more subjects’ data (87%) than the voxel-based approach (65%), generated more coherent tractography | Surface-based approach revealed more significant correlations between DTI metrics and five clinical scores |
| Tosun et al. (42) | 30 pts with CP only; 54 pts with epilepsy only; 38 pts with CP + epilepsy; 30 healthy children | BMD of lumbar vertebrae obtained by dual energy X-ray absorptiometry; clinical data (dietary Ca intake, whether intellectual disability, whether immobility, etc.) | Multivariate regression analysis used to evaluate the relationship between BMD and possible risk factors | Low BMD common in pts with CP and CP + epilepsy; risk factor of low BMD: immobility (not able to walk independently) | Low BMD related to the severity of CP, but not to vitamin D levels or AED treatment |
AHA, assisting hand assessment; BANIMM, bimanual activities negatively influenced by mirror movements; BMD, bone mineral density; BMI, body mass index; CST, corticospinal tract; CP, cerebral palsy; EMG, electromyographic; FA, fractional anisotropy; GMFCS, Gross Motor Function Classification System; HIE, hypoxia-ischemic encephalopathy; JTHFT, Jebsen taylor hand function test; MD, mean diffusivity; MMSE, multivariate multi-scale entropy; OPD, oropharyngeal dysphagia; PEL/VE, parenchymal echodensities/lucencies; Pts, patients; PVL, periventricular leukomalacia; ROI, region of interest; SDSC, Sleep Disturbance Scale for Children; STN, subthalamic nucleus; surgeon volume, the number of procedures performed; TMD, temporomandibular disorders; US, ultrasound.