Objectives: Perinatal stroke causes most hemiparetic cerebral palsy (CP). Pathophysiology is unknown with limited case-control studies. Comparisons to related disease states may provide unique insight. The population-based Canadian CP registry (CCPR) collects imaging and >70 risk factor variables. We hypothesized that perinatal strokes carry unique risk factor profiles within CP populations.

Methods: CCPR enrolls children across 13 Canadian centers. Systematic chart reviews and parental interviews acquire >150 variables including potential risk factors. Data current to June 2013 was extracted (RedCAP) and analyzed. Imaging reports for hemiparetic CP cases were classified as: (1) definitive stroke (arterial or venous), (2) probable stroke, (3) not stroke (alternate diagnosis), or (4) inconclusive. Risk factor variables were compared between definitive stroke and two disease control groups: (1) hemiparetic CP, not stroke and (2) all other CP. Univariate analysis informed multivariate logistic regression.

Results: Of 1168 children (57% male, median 42±23mos), hemiparetic was the most common CP phenotype (28%). Definitive perinatal stroke was common (158, 49%) compared to non-stroke (109, 34%) and 67% were arterial (33% venous). Comparing stroke to non-stroke hemiparetic CP found seven univariate associations but none persisted on multivariate analysis. Comparing stroke to other CP types demonstrated independent associations with preeclampsia (2.35;1.11-4.95;p=0.025), prematurity (0.165;0.093-0.293; p<0.001), maternal drug use (5.03;2.05-12.35;p<0.0001) and male gender (1.47;0.997-2.19;p=0.058).

Conclusions: CP registries are a valuable source of disease-specific etiologies including perinatal stroke. Stroke-specific risk factors may include previously (male gender) or inconsistently (preeclampsia) reported associations and novel factors (maternal drugs). Syndrome-controlled approaches may inform perinatal stroke pathogenesis.