We sought to determine the relative risk (RR) of Parkinson's disease (PD) for having a first-degree relative with PD versus having no first-degree relative with PD. Studies of familial aggregation of PD were identified by searching Medline and other sources. From each study, RRs were extracted or calculated based on the published data. Studies were categorized according to methodological characteristics, as well as by first-degree relationship type and age at PD onset restrictions. Meta-analyses and meta-regressions were based on random effect models. Twenty-nine studies of familial aggregation of PD were identified with results for first-degree relatives. The best estimate of the RR of PD for having a first-degree relative with PD was 2.9 (95% CI: 2.2, 3.8; P = 2.2 E-14), based on the studies with the most rigorous methods. The RR for sibling pairs was 4.4 (95% CI: 3.1, 6.1; P < 1.0 E-30), while for child-parent pairs it was 2.7 (95% CI: 2.0, 3.7; P = 3.6 E-10). The RR for early onset PD was 4.7 (95% CI: 3.2, 6.8; P = 6.7 E-16), while for late onset PD it was 2.7 (95% CI: 1.9, 3.9; P = 1.8 E-8). Inclusion of methodologically less rigorous investigations tended to increase the RR estimates. Summary RRs were clearly elevated above one for all study methods, all first-degree relationship types, and all age at onset categories. Familial aggregation of PD is strong and unlikely to be due to chance or to deficiencies in study methodology. © 2008 Movement Disorder Society.